ATM Switch Network Configuration Manual

ATM Switch Network
Configuration Manual
MANU0148-07
Revision B
11-17-1999
Software Version 6.1.x
Marconi plc
1000 FORE Drive
Warrendale, PA 15086-7502
Phone: 724-742-4444
FAX: 724-742-7742
http://www.marconi.com
Legal Notices
Copyright © 1995-1999 FORE Systems, Inc.
All rights reserved.
U.S. Government Restricted Rights. If you are licensing the Software on behalf of the U.S. Government (“Government”),
the following provisions apply to you. If the Software is supplied to the Department of Defense (“DoD”), it is classified as
“Commercial Computer Software” under paragraph 252.227-7014 of the DoD Supplement to the Federal Acquisition
Regulations (“DFARS”) (or any successor regulations) and the Government is acquiring only the license rights granted
herein (the license rights customarily provided to non-Government users). If the Software is supplied to any unit or agency
of the Government other than DoD, it is classified as “Restricted Computer Software” and the Government’s rights in the
Software are defined in paragraph 52.227-19 of the Federal Acquisition Regulations (“FAR”) (or any successor regulations)
or, in the cases of NASA, in paragraph 18.52.227-86 of the NASA Supplement to the FAR (or any successor regulations).
Restricted and Limited Rights Legend. Use, duplication or disclosure by the government is subject to restrictions as set
forth in subparagraph (c)(1)(ii) of the Rights in Technical Data and Computer Software clause at DFARS 252.227-7013
(October 1988) and FAR 52.227-19 (June 1987).
Printed in the USA.
No part of this work may be reproduced in any form.
The software and this publication is provided by FORE Systems, Inc. “as-is” without warranty of any kind, either express
or implied, including, but not limited to, the implied warranties of merchantability or fitness for a particular purpose.
FORE Systems, Inc. shall not be liable for any errors or omissions which may occur in the software or this publication, nor
for any direct, indirect, incidental, special, exemplary or consequential damages of any kind resulting from the furnishing,
performance or use of the software or this publication.
Information published here is current as of the date of publication of this document. Because FORE Systems, Inc. is
improving and adding features to its products continuously, the information in this document is subject to change without
notice.
The VxWorks software used in the Mini Loader is licensed from Wind River Systems, Inc., Copyright ©1984-1996.
TRADEMARKS
FORE Systems®, ForeRunner®, ForeRunnerLE®, ForeThought®, ForeView®, AVA®, CellPath®, Euristix®, Raceman®, and
Networks of Steel® are registered trademarks of FORE Systems, Inc.
ForeRunnerHE™, CellChain™, CellStarter™, PowerCell™, PowerHub™, ForeMan™, VoicePlus™, FramePlus™, StreamRun-
ner™, EdgeRunner™, ATV™, All Roads Lead To ATM™, ASN™, MSC™, TNX™, Intelligent Infrastructure™, I2™, Net-
pro™, Zero Hop Routing™, Application Aware™, ASX™, ESX™, NSX™, ForeWare™, ServiceGrid™, Dynamic Protection
Switching™, Capacity Aware Routing™, Demarc™, ERP Express™, ForeView Foundation™, ForeView Son™, ServiceOn
Management™, Active Switch™, ChannelFore™, UserAware™, Evolution™, AuthentiFirst Agent™, NetResource
Guard™, PriorSynch Solutions™, TAM™, and Technical Account Management™ are trademarks of FORE Systems, Inc.
Network of Steel® is a registered service mark of FORE Systems, Inc.
CellStarter SM and Networks of Steel SM are service marks of Fore Systems, Inc.
FCC CLASS A NOTICE

WARNING: Changes or modifications to this unit not expressly approved by the party responsible for compliance could
void this user’s authority to operate this equipment.
NOTE: The ASX™-200WG, the ASX-200BX, the ASX-1000, and the ForeRunnerLE® 155 have been tested and found to com-
ply with the limits for a Class A digital device, pursuant to Part 15 of the FCC Rules. These limits are designed to provide
reasonable protection against harmful interference when the equipment is operated in a commercial environment. This
equipment generates, uses, and can radiate radio frequency energy and, if not installed and used in accordance with the
instruction manual, may cause harmful interference to radio communications. Operation of the equipment in a residential
area is likely to cause harmful interference in which case the user will be required to correct the interference at his own
expense.
DOC CLASS A NOTICE

This digital apparatus does not exceed Class A limits for radio noise emission for a digital device as set out in the Radio
Interference Regulations of the Canadian Department of Communications.
Le present appareil numerique n’emet pas de bruits radioelectriques depassant les limites applicables aux appareils
numeriques de la class A prescrites dans le reglement sur le brouillage radioelectrique edicte par le ministere des
Communications du Canada.
NOTICE
Marking by the symbol indicates compliance of this system to the EMC (Electromagnetic Compatibility) directive of
the European Community and compliance to the Low Voltage (Safety) Directive. Such marking is indicative that this sys-
tem meets or exceeds the following technical standards:
• EN 55022 - “Limits and Methods of Measurement of Radio Interference Characteristics of Information
Technology Equipment.”
• EN 50082-1 - “Electromagnetic compatibility - Generic immunity standard Part 1: Residential, commercial,
and light industry.”
• IEC 1000-4-2 - “Electromagnetic compatibility for industrial-process measurement and control equipment
Part 2: Electrostatic discharge requirements.”
Part 3: Radiate electromagnetic field requirements.”
Part 4: Electrical fast transient/burst requirements.”
VCCI CLASS A NOTICE
This is a Class A product based on the standard of the Voluntary Control Council for Interference by Information
Technology Equipment (VCCI). If this equipment is used in a domestic environment, radio disturbance may arise. When
such trouble occurs, the user may be required to take corrective actions.
AUSTRALIA EMC COMPLIANCE
This product has been tested and found to comply with the Class A electromagnetic compatibility limits specified in AS/
NZ 3548.
FCC REQUIREMENTS (Notice to Users of DS1 Service)

The following instructions are provided to ensure compliance with the FCC Rules, Part 68.
(1) This device must only be connected to the DS1 network connected behind an FCC Part 68
registered channel service unit. Direct connection is not allowed.
(2) Before connecting your unit, you must inform the telephone company of the following
information:
Port ID REN/SOC FIC USOC

All DS1 Network Modules 6.0N 04DU9-BN/DN/1KN/1SN/1ZN RJ48C
(3) If the unit appears to be malfunctioning, it should be disconnected from the telephone lines
until you learn if your equipment or the telephone line is the source of the trouble. If your
equipment needs repair, it should not be reconnected until it is repaired.
(4) If the telephone company finds that this equipment is exceeding tolerable parameters, the
telephone company can temporarily disconnect service, although they will attempt to give
you advance notice if possible.
(5) Under the FCC Rules, no customer is authorized to repair this equipment. This restriction
applies regardless of whether the equipment is in or out of warranty.
(6) If the telephone company alters their equipment in a manner that will affect use of this
device, they must give you advance warning so as to give you the opportunity for
uninterrupted service. You will be advised of your right to file a complaint with the FCC.
(7) An affidavit is required to be given to the telephone company to affirm that no encoded ana-
log content or billing information is being transmitted.
CANADIAN IC CS-03 COMPLIANCE STATEMENT

NOTICE: The Industry Canada label identifies certified equipment. This certification means that the equipment meets
certain telecommunications network protective, operational and safety requirements. The Industry Canada label does not
guarantee the equipment will operate to the user’s satisfaction.
Before installing this equipment, users should ensure that it is permissible to be connected to the facilities of the local
telecommunications company. The equipment must also be installed using an acceptable method of connection. In some
cases, the company’s inside wiring associated with a single line individual service may be extended by means of a certified
connector assembly (telephone extension cord). The customer should be aware that compliance with the above conditions
may not prevent degradation of service in some situations.
Repairs to certified equipment should be made by an authorized Canadian maintenance facility designated by the sup-
plier. Any repairs or alterations made by the user to this equipment, or equipment malfunctions, may give the
telecommunications company cause to request the user to disconnect the equipment.
Users should ensure for their own protection that the electrical ground connections of the power utility, telephone lines and
internal metallic water pipe system, if present, are connected together. This precaution may be particularly important in
rural areas.
Caution: Users should not attempt to make such connections themselves, but should contact the appropriate electric
inspection authority, or electrician, as appropriate.
E1 AND E3 NOTICE
The E1 and E3 network modules that are described in this manual are approved for use in FORE Systems, Inc.’s host sys-
tems providing that the instructions below are strictly observed. Failure to follow these instructions invalidates the
approval.
Pan European Approval - Marking

Pan European approval of the E1 network module was issued by BABT following assessment against CTR12. This means
that it can be connected to ONP and unstructured PTO-provided private circuits with 120 Ω interfaces in all European
countries, according to Telecommunications Terminal Equipment (TTE) Directive 91/263/EEC. Thus, the following CE
mark applies:
168 X
The E1 and E3 network modules conform to safety standard EN60950 1992 following the provisions of Low Voltage
Product Safety Directive 73/23/EEC and CE Marking Directive 93/68/EEC, and can be marked accordingly with the CE
symbol.
The E1 and E3 network modules conform to EN55022 1994 and EN50082-1 1992 following the provisions of the EMC
Directive 89/336/EEC, and can be marked accordingly with the CE symbol.
National Approvals
UK
Network Module Connects to Approval Number

E1 Structured and unstructured AA60953
PTO-provided private circuits
with 75 Ω interfaces
E3 PTO-provided private circuits NS/4387/1/T/605954
with 75 Ω interfaces
Required User Guide Statements - UK Installation

The use of auxiliary products not authorized by FORE Systems, Inc. in FORE Systems, Inc. ATM Switches may cause the
power specification to be exceeded and is a potential safety hazard.
The equipment must be installed such that with the exception of the connections to the host, clearance and creepage
distances shown in the table below are maintained between the network module and any other assemblies which use or
generate a voltage shown in the table below. The larger distance shown in brackets applies where the local environment
within the host is subject to conductive pollution or dry non-conductive pollution which could become conductive due to
condensation. Failure to maintain these minimum distances invalidates the approval.
Voltage Used or Generated
Clearance (mm) Creepage (mm)
by Host or by Network Modules
2.0 2.4 (3.8) Up to 50 Vrms or Vdc
2.6 3.0 (4.8) Up to 125 Vrms or Vdc
4.0 5.0 (8.0) Up to 250 Vrms or Vdc
4.6 6.4 (10.0) Up to 300 Vrms or Vdc
For a host or other expansion card fitted in the host, using or generating voltages greater Above 300 Vrms or Vdc
than 300V (rms or dc), advice from a competent telecommunications engineer must be
obtained before installation of the relevant equipment.
NOTE: Installing the network modules in the appropriate FORE Systems hosts, according to the installation instructions
provided, satisfies the requirements listed above.
The following tables show the available ports and their safety status:
NM-8/E1D and NM-4/E1D
Ports Safety Status

E1 Ports TNV operating at SELV
Bus Connector SELV
NM-4/E3D
Ports Safety Status

E3 Ports TNV operating at SELV
Bus Connector SELV
SAFETY CERTIFICATIONS
ETL certified to meet Information Technology Equipment safety standards UL 1950, CSA 22.2 No. 950, and EN 60950.
ALPHA/BETA TEST DISCLAIMER

This ALPHA/BETA equipment has not been tested and found to comply with the emissions limits. These limits are
designed to provide reasonable protection against harmful interference. This equipment generates, uses, and can radiate
radio frequency energy and, if not installed and used in accordance with the instructions, may cause harmful interference
with radio communications. However, there is no guarantee that interference will not occur in a particular installation. If
this equipment does cause harmful interference to radio or television, which can be determined by turning the equipment
off and on, the user is encouraged to try to correct the interference by one or more of the following measures: 1) Reorient or
relocate the receiving antenna. 2) Increase the separation between the equipment and the receiver. 3) Connect the
equipment into an outlet on a circuit different from that to which the receiver is connected. 4) Consult FORE Systems, Inc.
for more help. This equipment is to be used for evaluation purposes. Changes must not be made without the prior proper
approval of FORE Systems, Inc.
This device has not been approved by the Federal Communications Commission. This device is not, and may not be,
offered for sale or lease, or sold or leased until the approval of the FCC has been obtained.
Linux - PAM
The software which is the subject matter of this manual incorporates certain files commonly known as Linux-PAM soft-
ware (“Linux-PAM”). FORE Systems, Inc. is not in any manner to be construed as the author of the Linux-PAM. The
authority of FORE Systems, Inc. to distribute the Linux-PAM is derived from the license contained in the following list of
conditions. FORE Systems, Inc. expressly disclaims any and all other right, title or interest in or to the Linux-PAM.
Any distributor of the Linux-PAM is required to reproduce the following copyright notices, list of conditions and dis-
claimer in any associated documentation. The following copyright notices, list of conditions and disclaimer are applicable
to the Linux-PAM only.
Copyright © Theodore Ts’o, 1996. All rights reserved.
Copyright © Andrew G. Morgan <[email protected]>, 1996-8.
Copyright © Theodore Ts’o, 1996. All rights reserved.
Copyright © Theodore Ts’o, 1996.
Copyright © Andrew Morgan, 1996-8.
Copyright © Andrew G. Morgan <[email protected]> 1996-8.
Copyright © Andrew G. Morgan <[email protected]> 1996,1997.
Copyright © 1995 by Red Hat Software, Marc Ewing.
Copyright © 1996-8, Andrew G. Morgan [email protected].
Copyright © 1995 by Red Hat Software, Michael K. Johnson.
Copyright © 1998, Andrew G. Morgan [email protected].
Redistribution and use in source and binary forms of Linux-PAM, with or without modification, are permitted provided
that the following conditions are met:
(1) Redistributions of source code must retain any existing copyright notice, and this entire per-
mission notice in its entirety, including the disclaimer of warranties.
(2) Redistributions in binary form must reproduce all prior and current copyright notices, this
list of conditions and the following disclaimer in the documentation and/or other materials
provided with the distribution.
(3) The name of any author may not be used to endorse or promote products derived from this
software without their specific prior written permission.
ALTERNATIVELY, this product may be distributed under the terms of the GNU General Public License, in which case the
provisions of the GNU GPL are required INSTEAD OF the above restrictions. (This clause is necessary due to a potential
conflict between the GNU GPL and the restrictions contained in a BSD-style copyright.)
THIS SOFTWARE IS PROVIDED “AS IS”' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT
OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOW-
EVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN
IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Table of Contents
List of Figures
List of Tables
Preface
Chapter Summaries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i
Related Manuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ii
Technical Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Typographical Styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
Important Information Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v
Laser Warning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
Safety Precautions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Modifications to Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Placement of a Marconi Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Power Cord Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
CHAPTER 1 Configuring PVCs

1.1 General Concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 1
1.2 Virtual Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 3
1.2.1 Through Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 4
1.2.2 Originating and Terminating Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 6
1.3 Listing Virtual Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 7
1.3.1 Listing Through Paths. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 7
1.3.2 Listing Originating and Terminating Paths . . . . . . . . . . . . . . . . . . . . . . . 1 - 9
1.4 Virtual Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 11
1.4.1 Smart Permanent Virtual Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 13
1.4.2 Listing Virtual Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 15
1.5 Creating PVCs and SPVCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 17
1.5.1 Creating a Through Path . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 18
1.5.2 Creating an Originating or Terminating Path . . . . . . . . . . . . . . . . . . . . 1 - 20
1.5.2.1 Shaping Multiple Originating Paths on a Single Port. . . . . . . . 1 - 22
1.5.2.2 Terminating a PVC at a Switch . . . . . . . . . . . . . . . . . . . . . . . . 1 - 26
1.5.2.3 Creating ATM ARP Entries . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 26
1.5.2.4 Listing ATM ARP Entries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 27
1.5.3 Creating a Virtual Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 28
ATM Switch Network Configuration Manual TOC - 1

Table of Contents
1.5.4 Creating SPVCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 30

1.5.4.1 Creating a SPANS SPVC . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 30
1.5.4.2 Creating a PP PNNI SPVC. . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 33
1.5.4.3 Creating a PMP PNNI SPVC . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 35
1.6 Traffic Types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 38
1.7 Traffic Policing (Usage Parameter Control) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 39
1.7.1 Leaky Bucket Algorithm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 39
1.7.2 Non-conforming Cells: Tagging vs. Dropping . . . . . . . . . . . . . . . . . . . 1 - 40
1.7.3 UPC Traffic Contract Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 41
1.7.4 AMI UPC Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 42
CHAPTER 2 Configuring Classical IP

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 1
2.1.1 Logical IP Subnets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 2
2.1.2 Classical IP Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 2
2.1.3 SPANS Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 3
2.2 Address Registration and ILMI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 4
2.2.1 NSAP Addresses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 4
2.2.2 Operating with ILMI Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 5
2.2.3 Operating without ILMI Support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 5
2.2.4 Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 5
2.3 ARP and ARP Servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 6
2.3.1 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 6
2.3.2 Configuring a FORE Switch to be an ARP Server . . . . . . . . . . . . . . . . 2 - 7
2.3.3 Classical IP Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 8
2.3.4 Operational Issues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 9
2.4 Classical IP PVCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 10
2.4.1 Theory and Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 10
2.4.2 Revalidation and Removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 10
2.5 Configuring the Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 11
2.5.1 Third-Party Host with No ILMI and No RFC-1577 Support . . . . . . . . . 2 - 12
2.5.2 Third-Party Switch with ILMI and No RFC-1577 Support . . . . . . . . . . 2 - 13
2.5.3 Third-Party Switch with RFC-1577 and No ILMI Support . . . . . . . . . . 2 - 14
CHAPTER 3 Configuring an Emulated LAN

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.1.1 Ethernet ELANs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.1.2 Token Ring ELANs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
3.2 ELAN Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2
3.2.1 LAN Emulation Client (LEC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
3.2.2 LAN Emulation Configuration Server (LECS) . . . . . . . . . . . . . . . . . . . . 3-3
TOC - 2 ATM Switch Network Configuration Manual

Table of Contents
3.2.3 LAN Emulation Server (LES) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3

3.2.4 Broadcast and Unknown Server (BUS) . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3
3.3 Emulated LAN Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 4
3.3.1 Initialization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 6
3.3.2 Registration and Address Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 7
3.3.3 Data Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 8
3.4 Distributed LAN Emulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 9
3.4.1 Single Server LANE Services Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 9
3.4.1.1 Using a Single Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 10
3.4.1.2 Limitations of a Single Server . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 11
3.4.2 Distributed LAN Emulation Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 12
3.4.2.1 Using DLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 13
3.4.2.2 Advantages of DLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 15
3.4.2.2.1 Load Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 15
3.4.2.2.2 Improved Performance for Remote LECs . . . . . . . . 3 - 15
3.4.2.2.3 Fault Tolerance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 15
3.4.2.2.3.1 Single Server ELAN. . . . . . . . . . . . . . . . 3 - 16
3.4.2.2.3.2 DLE ELAN . . . . . . . . . . . . . . . . . . . . . . . 3 - 18
3.5 ELAN Access Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 21
3.6 Configuring an ELAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 22
3.6.1 Configuring an LECS Configuration Database File . . . . . . . . . . . . . . . 3 - 24
3.6.1.1 Before You Begin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 24
3.6.1.2 LECS Configuration File Syntax . . . . . . . . . . . . . . . . . . . . . . . 3 - 25
3.6.1.3 Defining an ELAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 31
3.6.1.4 Defining a Client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 32
3.6.1.5 LECS Control Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 33
3.6.1.6 LECS MPOA Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 34
3.6.2 Sample LECS Configuration File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 36
3.6.3 Starting the LAN Emulation Services . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 39
3.6.3.1 Starting the LECS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 39
3.6.3.2 Starting the DLE LES/BUS Peer Servers . . . . . . . . . . . . . . . . 3 - 41
3.6.4 Starting the LEC(s) and Joining an ELAN . . . . . . . . . . . . . . . . . . . . . . 3 - 44
3.7 Upgrading an ELAN to Use DLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 47
3.7.1 Edit the LECS.CFG File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 48
3.7.2 Delete the LES and BUS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 49
3.7.3 Upgrade the Switches Running Services . . . . . . . . . . . . . . . . . . . . . . . 3 - 50
3.7.4 Create the DLE Peer Servers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 50
3.7.5 Transfer the Updated LECS.CFG File . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 52
3.7.6 Restart the LECS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 52
3.7.7 Recreate the LECs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 52
3.7.8 Create the Last DLE Peer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 54
3.7.9 Add the Last DLE Peer to Each Peer List. . . . . . . . . . . . . . . . . . . . . . . 3 - 54

Table of Contents
3.7.10 Update the LECS.CFG File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 55

3.7.11 Transfer the Final LECS.CFG File . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 55
3.7.12 Restart the LECS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 56
3.8 Upgrading an ELAN without Using DLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 57
3.8.1 Deleting the Non Co-located Services . . . . . . . . . . . . . . . . . . . . . . . . 3 - 58
3.8.1.1 Administer Down the Services . . . . . . . . . . . . . . . . . . . . . . . . 3 - 58
3.8.1.2 Delete the Non Co-located LES and BUS . . . . . . . . . . . . . . . 3 - 58
3.8.1.2.1 Edit the LECS.CFG File. . . . . . . . . . . . . . . . . . . . . 3 - 58
3.8.2 Upgrade the Switches Running Services . . . . . . . . . . . . . . . . . . . . . . 3 - 59
3.8.3 Recreate the LES and BUS Together . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 59
3.8.4 Administer the Services Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 60
CHAPTER 4 MPOA
4.1 Overview of LANE/MPOA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 1
4.2 LANE Primer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 2
4.2.1 LANE Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 2
4.2.2 An Example LANE Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 3
4.2.2.1 The Initialization Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 4
4.2.2.2 The Connection Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 5
4.2.2.3 Multicast and Broadcast Packets . . . . . . . . . . . . . . . . . . . . . . . 4 - 5
4.2.2.4 Accessing Fast Ethernet and FDDI Networks . . . . . . . . . . . . . 4 - 5
4.2.2.5 Multiple ELANs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 6
4.2.2.6 Distributed LAN Emulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 6
4.2.2.7 Automatic ELAN Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 6
4.2.2.8 Intelligent BUS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 6
4.3 An Introduction to Multi-Protocol Over ATM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 7
4.3.1 LANE Without MPOA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 7
4.3.2 Why MPOA? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 8
4.3.3 MPOA Components. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 10
4.3.4 MPOA Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 11
4.3.4.1 MPS Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 12
4.3.4.2 Initialization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 12
4.3.4.3 Flow Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 13
4.3.4.4 Making a Shortcut . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 13
4.3.4.5 Shortcut Teardown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 14
CHAPTER 5 ForeThought PNNI

5.1 FT-PNNI Routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
5.1.1 Hello Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
5.1.2 Topology Database Exchange. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
5.1.3 Flooding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
5.1.4 Hierarchical Routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3

Table of Contents
5.1.4.1
Hierarchical Addressing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 3
5.1.4.1.1 Switch Prefix. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 3
5.1.4.1.2 Switch Summary Prefix . . . . . . . . . . . . . . . . . . . . . . 5 - 4
5.1.4.1.3 Peer Group ID. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 4
5.1.4.2 Path Computation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 4
5.2 The Physical Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 5
5.2.1 Peer Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 8
5.2.2 Peer Group Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 8
5.2.3 Border Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 8
5.2.4 Peer Group Summary Node (PGSN) . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 9
5.2.5 Backbone Topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 9
5.2.6 Single Switch Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 9
CHAPTER 6 ATM Forum PNNI

6.1 PNNI Routing Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 2
6.1.1 Hello Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 2
6.1.2 Database Exchange Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 2
6.1.3 Flooding Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 3
6.1.4 Path Computation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 3
6.1.5 Hierarchical Routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 4
6.1.5.1 Peer Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 5
6.1.5.2 PGL/LGN Hierarchy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 6
6.1.5.3 Border Nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 7
6.1.5.4 Lop-sided Hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 7
6.2 PNNI Signalling Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 8
6.2.1 Source Routing and DTLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 8
6.2.2 Crankback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 9
6.3 Border Node, PGL, and LGN Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 10
6.3.1 Outside Links and Hello Protocol over Outside Links. . . . . . . . . . . . . . 6 - 11
6.3.2 Origination of Uplinks and Uplink PTSEs . . . . . . . . . . . . . . . . . . . . . . . 6 - 11
6.3.3 Entry and Exit Border Node DTL Handling. . . . . . . . . . . . . . . . . . . . . . 6 - 11
6.3.4 Entry Border Node Crankback Handling . . . . . . . . . . . . . . . . . . . . . . . 6 - 11
6.3.5 SVCC RCC Establishment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 12
6.3.6 SVCC RCC Hello Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 12
6.3.7 LGN Horizontal Link Hello Protocol and PTSEs. . . . . . . . . . . . . . . . . . 6 - 12
6.3.8 PGLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 13
6.3.8.1 Parent Nodes and Child Nodes . . . . . . . . . . . . . . . . . . . . . . . . 6 - 13
6.3.8.2 PGL Election . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 13
6.3.8.3 PGL Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 13
6.3.9 Aggregation Tokens and Link Aggregation. . . . . . . . . . . . . . . . . . . . . . 6 - 14
6.3.9.1 Aggregation Tokens. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 14
6.3.9.2 Derived Aggregation Tokens . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 14

Table of Contents
6.3.9.3 Link Aggregation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 15

6.3.10 Nodal Aggregation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 15
6.3.11 Internal and Exterior Address Summarization. . . . . . . . . . . . . . . . . . . 6 - 15
6.3.12 Anycast Address Scoping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 16
6.3.12.1 Organizational Scope and PNNI Scope . . . . . . . . . . . . . . . . . 6 - 16
6.3.12.2 Connection Scope Selection . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 16
6.4 Internetworking between PNNI and FT-PNNI . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 17
6.4.1 Gateway Switches and Split Switches. . . . . . . . . . . . . . . . . . . . . . . . . 6 - 17
6.4.1.1 Using ForeThought 5.x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 17
6.4.1.2 Using ForeThought 6.x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 18
6.4.2 Interfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 18
6.4.3 Dynamic Leaking of Reachability Information . . . . . . . . . . . . . . . . . . . 6 - 18
6.4.3.1 Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 19
6.4.3.1.1 Peer Groups in Areas . . . . . . . . . . . . . . . . . . . . . . 6 - 20
6.4.3.1.2 Area IDs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 20
6.4.3.1.3 Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 21
6.4.3.2 Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 21
6.4.3.2.1 Configuring Domains . . . . . . . . . . . . . . . . . . . . . . . 6 - 22
6.4.3.3 Propagation of Reachability Information. . . . . . . . . . . . . . . . . 6 - 23
6.4.3.3.1 Policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 23
6.4.3.3.1.1 Policies in a Flat PNNI Network . . . . . . 6 - 23
6.4.3.3.1.2 Policies in a Hierarchical PNNI
Network. . . . . . . . . . . . . . . . . . . . . . . . . 6 - 23
6.4.3.3.2 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 24
6.4.3.3.2.1 Scope in a Flat PNNI Network . . . . . . . 6 - 24
6.4.3.3.2.2 Scope in a Hierarchical PNNI
Network. . . . . . . . . . . . . . . . . . . . . . . . . 6 - 24
6.4.3.3.3 The Process for Leaking Reachability
Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 24
6.5 Path Extension QoS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 25
6.6 Load-balanced Routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 26
6.7 Load-balanced (Congestion-based) UBR Routing . . . . . . . . . . . . . . . . . . . . . . . 6 - 27
6.7.1 Advertising Congestion Metrics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 27
6.7.2 Routing Using Congestion Information . . . . . . . . . . . . . . . . . . . . . . . . 6 - 28
6.7.3 Enabling Congestion-based UBR Routing . . . . . . . . . . . . . . . . . . . . . 6 - 29
6.7.4 Other Parameters Affecting Congestion-based UBR . . . . . . . . . . . . . 6 - 29
6.7.5 Load-balanced Routing vs. Congestion-based Routing . . . . . . . . . . . 6 - 30
6.8 Configuring PNNI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 31
6.8.1 Doing the Preparation Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 31
6.8.1.1 Dividing the NSAP Address Space. . . . . . . . . . . . . . . . . . . . . 6 - 32
6.8.1.2 Address Summarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 32
6.8.1.3 Determining the Layers of Hierarchy Needed. . . . . . . . . . . . . 6 - 32

Table of Contents
6.8.1.3.1 Enterprise Example . . . . . . . . . . . . . . . . . . . . . . . . 6 - 33

6.8.1.3.2 Service Provider Example . . . . . . . . . . . . . . . . . . . 6 - 34
6.8.1.3.3 Configuration Example . . . . . . . . . . . . . . . . . . . . . . 6 - 35
6.8.2 Configuring the NSAP Addresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 35
6.8.3 Choosing a PGL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 37
6.8.4 Configuring the Parent Nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 37
6.8.5 Configuring the PGL Priority . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 38
6.8.6 Configuring the Remaining Peer Groups . . . . . . . . . . . . . . . . . . . . . . . 6 - 40
6.8.7 Verifying That the Hierarchy Has Come Up . . . . . . . . . . . . . . . . . . . . . 6 - 42
CHAPTER 7 Signalling
7.1 UNI 4.0 Supplementary Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 1
7.1.1 Description of Supplementary Services . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 2
7.1.1.1 Direct Dialing In (DDI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 3
7.1.1.2 Multiple Subscriber Number (MSN). . . . . . . . . . . . . . . . . . . . . . 7 - 3
7.1.1.3 Calling Line Identification Presentation (CLIP) . . . . . . . . . . . . . 7 - 4
7.1.1.4 Calling Line Identification Restriction (CLIR) . . . . . . . . . . . . . . . 7 - 4
7.1.1.5 Connected Line Identification Presentation (COLP) . . . . . . . . . 7 - 4
7.1.1.6 Connected Line Identification restriction (COLR) . . . . . . . . . . . 7 - 5
7.1.1.7 Sub Addressing (SUB) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 5
7.1.1.8 User-to-User Signalling (UUS) . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 5
7.1.2 Configuring Supplementary Services. . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 6
7.2 Virtual UNI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 8
7.3 Proxy Signalling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 14
7.4 VCI Allocation Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 18
7.4.1 Determining the VCI Allocation Range with ILMI Down . . . . . . . . . . . . 7 - 18
7.4.2 Determining the VCI Allocation Range with ILMI Up . . . . . . . . . . . . . . 7 - 20
7.5 Signalling Scope. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 23
7.5.1 VC-Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 23
7.5.2 Dynamic Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 24
7.6 Signalling Channel Auto Configuration Procedures . . . . . . . . . . . . . . . . . . . . . . . 7 - 25
7.6.1 Overview of Signalling Channel Auto Configuration . . . . . . . . . . . . . . . 7 - 25
7.6.2 Rules for Signalling Channel Auto Configuration . . . . . . . . . . . . . . . . . 7 - 28
7.6.2.1 Specifying the Type and Interface Version. . . . . . . . . . . . . . . . 7 - 28
7.6.2.1.1 Examples of Valid Configurations . . . . . . . . . . . . . . 7 - 29
7.6.2.1.2 Examples of Invalid Configurations . . . . . . . . . . . . 7 - 30
7.6.2.2 Specifying the Scope and Mode . . . . . . . . . . . . . . . . . . . . . . . 7 - 31
7.6.2.2.1 Examples of Valid Configurations . . . . . . . . . . . . . . 7 - 32
7.6.2.2.2 Examples of Invalid Configurations . . . . . . . . . . . . 7 - 33
7.6.2.3 Specifying the User/Network Side . . . . . . . . . . . . . . . . . . . . . . 7 - 34
7.7 Allowable Combination of Traffic Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 35

Table of Contents
7.7.1 PNNI 1.0/UNI 4.0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 35

7.7.1.1 Service Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 35
7.7.1.2 Allowable Combination of Traffic Parameters . . . . . . . . . . . . . 7 - 36
7.7.2 UNI 3.X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 36
CHAPTER 8 Security
8.1 Configuring Userids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 1
8.1.1 Login Authentication Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 5
8.1.1.1 Local Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 5
8.1.1.2 SecurID Authentication. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 5
8.1.1.2.1 SecurID Protection on Switches . . . . . . . . . . . . . . . 8 - 5
8.1.1.2.2 SecurID Passcode . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 6
8.1.1.2.2.1 PIN Number . . . . . . . . . . . . . . . . . . . . . . 8 - 6
8.1.1.2.2.2 SecurID Tokens. . . . . . . . . . . . . . . . . . . . 8 - 6
8.1.1.2.3 SecurID Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 6
8.1.1.2.3.1 Slave Server . . . . . . . . . . . . . . . . . . . . . . 8 - 6
8.1.1.2.3.2 Server Database. . . . . . . . . . . . . . . . . . . 8 - 7
8.1.1.2.3.3 Data Encryption between the Server
and Switches . . . . . . . . . . . . . . . . . . . . . 8 - 7
8.1.1.2.4 SecurID AMI Commands . . . . . . . . . . . . . . . . . . . . . 8 - 7
8.1.1.2.5 Configuring SecurID on a Switch. . . . . . . . . . . . . . . 8 - 7
8.1.1.2.5.1 Installing the Server Software . . . . . . . . . 8 - 7
8.1.1.2.5.2 Transferring the Configuration File . . . . . 8 - 7
8.1.1.2.5.3 Editing the Server Configuration File . . . 8 - 8
8.1.1.2.5.4 An Example Login Using SecurID . . . . 8 - 10
8.1.1.3 Kerberos Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 11
8.1.1.3.1 Kerberos Server. . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 11
8.1.1.3.2 Kerberos AMI Commands . . . . . . . . . . . . . . . . . . . 8 - 11
8.1.1.3.2.1 Configuring Kerberos Authentication
on a Switch . . . . . . . . . . . . . . . . . . . . . . 8 - 12
8.1.1.3.2.2 Installing the Kerberos Server
Software . . . . . . . . . . . . . . . . . . . . . . . . 8 - 12
8.1.1.3.2.3 Configuring the Kerberos Server
Database . . . . . . . . . . . . . . . . . . . . . . . 8 - 12
8.1.1.3.2.4 Transferring the Secret Key File . . . . . . 8 - 12
8.1.1.3.2.5 Setting the Correct Time . . . . . . . . . . . . 8 - 13
8.1.1.3.2.6 Adding IP Address-to-Hostname
Mappings . . . . . . . . . . . . . . . . . . . . . . . 8 - 13
8.1.1.3.2.7 Establish an Encrypted Telnet
Session . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 14
8.1.1.3.3 An Example Login Using Kerberos
Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 14

Table of Contents
8.1.2 AMI Command Privileges. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 15

8.1.2.1 Admin Privileges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 15
8.1.2.2 User Privileges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 15
8.1.2.3 Readonly Privileges. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 15
8.1.3 AMI Access Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 16
8.1.4 Userid Password . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 16
8.1.5 Privilege Level for Unlisted Users. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 16
8.2 IP Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 17
8.2.1 Authorized IP Address Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 17
8.2.2 IP Filtering Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 17
8.2.2.1 Strict Source Routing Flag . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 18
8.2.2.2 Loose Source Routing Flag . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 18
8.2.2.3 All Flag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 18
8.2.3 IP Access Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 18
8.3 NSAP Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 19
8.3.1 Filters and Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 19
8.3.2 NSAP Filtering Lookup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 20
8.3.3 NSAP Filtering Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 21
CHAPTER 9 Configuring Timing

9.1 Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 1
9.2 Timing Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 1
9.3 Switchclock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 2
9.3.1 Failover of the Switchclock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 2
9.4 Port Level Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 3
9.5 Timing Configuration Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 4
9.5.1 Configuring Timing on an ASX-200BX, TNX-210, or ASX-200WG . . . . 9 - 4
9.5.2 Configuring Timing on an ASX-1000, ASX-1200, or TNX-1100
(Single Timing Domain) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 4
(Multiple Timing Domains) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 5
9.5.4 Configuring Timing on a ForeRunnerLE 155 or LE 25 . . . . . . . . . . . . . . 9 - 5
9.5.5 Configuring Timing on an ASX-4000 . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 6
9.5.6 Configuring Timing on an ESX-3000 . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 6
APPENDIX A Configuring SNMP

A.1 SNMP Indexing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .A - 1
A.2 SNMP Traps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .A - 2
A.2.1 Adding SNMP Trap Destinations . . . . . . . . . . . . . . . . . . . . . . . . . . . . .A - 24
A.2.2 Displaying SNMP Trap Destinations. . . . . . . . . . . . . . . . . . . . . . . . . . .A - 24
A.2.3 Removing SNMP Trap Destinations . . . . . . . . . . . . . . . . . . . . . . . . . . .A - 25

Table of Contents
APPENDIX B Configuring Circuit Emulation Services

B.1 Idle Channel Suppression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 2
B.1.1 CAS Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 2
B.1.1.1 Idle Detection using Signalling Bits . . . . . . . . . . . . . . . . . . . . . B - 2
B.1.1.2 Active Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 3
B.1.2 Basic Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 3
B.1.2.1 Idle Detection using Pattern Matching . . . . . . . . . . . . . . . . . . . B - 3
B.1.2.2 Idle Detection using Mask Patterns . . . . . . . . . . . . . . . . . . . . . B - 3
B.1.2.3 Active Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 4
B.1.3 Idle Channel Suppression Configuration Examples . . . . . . . . . . . . . . . B - 4
B.1.3.1 Enabling Idle Channel Suppression . . . . . . . . . . . . . . . . . . . . . B - 4
B.1.3.2 Configuring Idle Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 5
B.1.3.2.1 Configuring Idle CAS Signalling Pattern
Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 5
B.1.3.2.2 Configuring an Idle Pattern for Pattern
Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 6
B.1.3.2.3 Configuring Mask Patterns . . . . . . . . . . . . . . . . . . . B - 6
B.1.3.3 Configuring an Idle Integration Period . . . . . . . . . . . . . . . . . . . B - 7
B.2 Configuring CES Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 8
B.2.1 Deleting a CES Connection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 8
B.2.2 Creating a New CES Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 9
B.2.3 Displaying CES Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 15
B.2.3.1 Displaying Advanced Information . . . . . . . . . . . . . . . . . . . . . . B - 16
B.2.3.2 Displaying Idle Channel Suppression Information . . . . . . . . . B - 17
B.2.4 Displaying CES Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 18
APPENDIX C Converting FT-PNNI and PNNI Networks

C.1 ASX-1000, ASX-1200, and TNX-1100 Routing Configuration Issues . . . . . . . . . . C - 2
C.1.1 ASX-1000s, ASX-1200s, or TNX-1100s in FT-PNNI Peer Groups . . . . C - 2
C.1.2 ASX-1000s, ASX-1200s, or TNX-1100s in PNNI Areas . . . . . . . . . . . . C - 4
C.1.3 Multiple Gateways in an ASX-1000, or TNX-1100. . . . . . . . . . . . . . . . . C - 5
C.1.4 Migrating from FT-PNNI to PNNI Routing . . . . . . . . . . . . . . . . . . . . . . . C - 5
C.2 Migration of a Non-Hierarchical FT-PNNI Network . . . . . . . . . . . . . . . . . . . . . . . . C - 6
C.2.1 Migration Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 6
C.2.2 Detailed Migration Example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 7
C.3 Migration of a Hierarchical FT-PNNI Network . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 11
C.3.1 Migration of a Hierarchical FT-PNNI Network with a Contiguous
Backbone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 11
C.3.1.1 Migration Starting with the Backbone . . . . . . . . . . . . . . . . . . . C - 11
C.3.1.1.1 Migration Overview . . . . . . . . . . . . . . . . . . . . . . . . C - 12
C.3.1.1.1.1 Upgrade the Switches. . . . . . . . . . . . . . C - 12
C.3.1.1.1.2 Convert the Backbone . . . . . . . . . . . . . C - 12

Table of Contents
C.3.1.1.1.3 Convert the Individual Peer

Groups . . . . . . . . . . . . . . . . . . . . . . . . . .C - 16
C.3.1.2 Migration Starting with the Peer Groups . . . . . . . . . . . . . . . . .C - 21
C.3.1.2.1 Overview of the Migration . . . . . . . . . . . . . . . . . . . .C - 21
C.3.1.2.1.1 Upgrade the Switches . . . . . . . . . . . . . .C - 21
C.3.1.2.1.2 Convert Peer Group C . . . . . . . . . . . . . .C - 21
C.3.1.2.1.3 Convert Peer Group A . . . . . . . . . . . . . .C - 25
C.3.1.2.1.4 Convert the Backbone . . . . . . . . . . . . . .C - 25
C.3.1.2.1.5 Convert Peer Group B . . . . . . . . . . . . . .C - 28
C.3.2 Migration of a Hierarchical FT-PNNI Network with a
Non-Contiguous Backbone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .C - 29
C.4 Migration of a Non-hierarchical PNNI Network to a Lop-sided Hierarchical
PNNI Network. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .C - 31
C.4.1 Migration Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .C - 31
C.4.2 Detailed Migration Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .C - 32
APPENDIX D Configuring FramePlus Modules

D.1 Frame Relay Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 2
D.1.1 Interworking Function (IWF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 2
D.1.1.1 Translation Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 2
D.1.1.2 Transparent Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 2
D.2 Configuring the Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 3
D.2.1 Dividing the Buffer Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 3
D.2.2 Setting the Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 4
D.2.2.1 Noting the CLP0PPD Threshold . . . . . . . . . . . . . . . . . . . . . . . .D - 5
D.2.2.2 Configuring the CLP1EPD Threshold . . . . . . . . . . . . . . . . . . . .D - 6
D.2.2.3 Configuring the CLP0EPD Threshold . . . . . . . . . . . . . . . . . . . .D - 7
D.2.2.4 Configuring the CLP1PPD Threshold . . . . . . . . . . . . . . . . . . . .D - 8
D.3 Profiles and Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 9
D.3.1 EPD/PPD Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 9
D.3.2 FRF.8 Profile. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 9
D.3.3 Frame Relay Rate Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 10
D.3.4 LMI Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 11
D.3.5 Service Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 11
D.3.6 FUNI Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 11
D.3.7 Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 11
D.4 Configuring Frame Relay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 12
D.4.1 Choosing Frame Relay Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 12
D.4.2 Creating Frame Relay Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 13
D.4.2.1 Creating an Unstructured E1 Service . . . . . . . . . . . . . . . . . . .D - 14
D.4.3 Creating Frame Relay PVCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 15
D.4.4 Configuring Frame Relay SPVCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 17

Table of Contents
D.4.4.1 Creating a Frame Relay SPANS SPVC . . . . . . . . . . . . . . . . . D - 17

D.4.4.2 Creating a Frame Relay PNNI SPVC . . . . . . . . . . . . . . . . . . . D - 20
D.5 Configuring FUNI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D - 22
D.5.1 Changing the Application Key . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D - 22
D.5.2 Creating the Profiles for FUNI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D - 23
D.5.3 Creating FUNI Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D - 23
D.5.3.1 Creating an Unstructured E1 Service. . . . . . . . . . . . . . . . . . . D - 24
D.5.4 Creating FUNI PVCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D - 25
D.5.5 Configuring FUNI SPVCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D - 26
D.5.5.1 Creating a FUNI SPANS SPVC . . . . . . . . . . . . . . . . . . . . . . . D - 26
D.5.5.2 Creating a FUNI PNNI SPVC . . . . . . . . . . . . . . . . . . . . . . . . . D - 29
D.6 Upgrading the FramePlus Network Module Software . . . . . . . . . . . . . . . . . . . . . D - 31
APPENDIX E RMON
E.1 MIB Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-2
E.2 SNMP Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-3
E.3 RMON Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-4
E.3.1 RMON Table Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-4
E.3.1.1 The portSelect Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-5
E.3.1.2 The atmStats Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-6
E.3.1.3 The RMON Alarm Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-7
E.3.1.4 The Event Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E-8
APPENDIX F Configuring IMA Modules

F.1 IMA Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 2
F.1.1 Links and Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 2
F.1.2 Benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 3
F.1.2.1 Scalability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 3
F.1.2.2 Load-balancing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 3
F.1.2.3 ATM Connection Aware IMA. . . . . . . . . . . . . . . . . . . . . . . . . . . F - 3
F.1.2.4 Full Series D Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 3
F.2 Configuring IMA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 4
F.2.1 Creating a Group and Assigning Links . . . . . . . . . . . . . . . . . . . . . . . . . F - 4
F.2.1.1 Assigning Additional Links to the Group. . . . . . . . . . . . . . . . . . F - 5
F.2.1.2 Deleting Links from the Group . . . . . . . . . . . . . . . . . . . . . . . . . F - 6
F.2.2 Creating PVPs, PVCs, or SPVXs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 8
F.3 Testing an IMA Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 10
F.3.1 Choosing a Test Link . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 10
F.3.2 Choosing a Test Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 10
F.3.3 Starting or Stopping a Test Procedure . . . . . . . . . . . . . . . . . . . . . . . . F - 10
F.3.4 Test Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 11

Table of Contents
F.4 Upgrading the IMA Network Module Software . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 12
APPENDIX G ASX-4000 Redundancy

G.1 Overview of SONET APS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 2
G.1.1 Line/Port Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 2
G.1.1.1 Linear APS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 2
G.1.1.2 Unidirectional Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 3
G.1.1.3 Failure Recovery / Switchover Performance . . . . . . . . . . . . . . G - 3
G.1.1.4 Manual and Automatic Protection . . . . . . . . . . . . . . . . . . . . . . G - 3
G.1.1.5 Protection Failover Performance . . . . . . . . . . . . . . . . . . . . . . . G - 3
G.1.1.6 Configuration and Management . . . . . . . . . . . . . . . . . . . . . . . G - 4
G.1.2 Fabric and Port Card Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 6
G.1.2.1 Manual and Active Protection . . . . . . . . . . . . . . . . . . . . . . . . . G - 7
G.1.3 Timing Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 7
G.2 Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 8
G.3 Redundant Switch Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 9
G.3.1 20 Gbps Fully Redundant Configuration . . . . . . . . . . . . . . . . . . . . . . . G - 9
G.3.2 30 Gbps Mixed Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 10
G.3.3 10 Gbps Fully Redundant Configuration . . . . . . . . . . . . . . . . . . . . . . G - 11
G.3.4 Line Level and Port Card Redundancy Without Fabric
Redundancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 12
G.3.5 Fabric Redundancy Without Line Level and Port Card
Redundancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 13
G.4 Redundant Link Naming Convention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 14
G.5 Configuring Working and Protection Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . G - 15
G.6 Partner Synchronization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 20
G.6.1 SONET Port Level Commands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 20
G.6.2 Netmod Level Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 21
G.6.2.1 Series 1 Port Cards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 21
G.6.2.2 Series 2 Port Cards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 22
G.6.2.3 Buffer Class and Buffer Class Assignment . . . . . . . . . . . . . . G - 23
G.6.3 Fabric Level Commands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 24
G.6.4 Channelized OC-12 Port Card Commands . . . . . . . . . . . . . . . . . . . . G - 25
G.6.4.1 SONET Medium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 25
G.6.4.2 SONET Path . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 26
G.6.4.3 DSX3 Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 27
G.6.4.4 ATM Transmission Convergence . . . . . . . . . . . . . . . . . . . . . . G - 28
G.6.5 LED Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 29
G.6.6 Netmod Alarms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 29
G.7 Loopbacks on Redundant Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 30
Index

Table of Contents

List of Figures

Figure 1.1 The Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 2
Figure 1.2 Virtual Channels in a Virtual Path. . . . . . . . . . . . . . . . . . . . . . . . . 1 - 3
Figure 1.3 An Example of a Virtual Path . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 3
Figure 1.4 Composition of a Virtual Path . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 4
Figure 1.5 An Example of a Through Path. . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 5
Figure 1.6 Through Paths are Unidirectional . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 5
Figure 1.7 Using Originating and Terminating Paths for Bandwidth
Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 6
Figure 1.8 An Example of a Virtual Channel . . . . . . . . . . . . . . . . . . . . . . . . 1 - 11
Figure 1.9 Example of a Virtual Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 11
Figure 1.10 Virtual Channels are Unidirectional . . . . . . . . . . . . . . . . . . . . . . 1 - 12
Figure 1.11 Virtual Channels Created on Terminating Path C3|3 . . . . . . . . . 1 - 12
Figure 1.12 Virtual Channels Created on Originating Path C2|2 . . . . . . . . . . 1 - 13
Figure 1.13 The Path of a Cell Via SPVCs . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 14
Figure 1.14 PVPs Looped through Port 1A2 and Output on Port 1A1
to WAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 22
Figure 1.15 PVPs Coming in Port 1A1 from WAN and Looped through
Port 1A2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 23

Figure 2.1 Configuring a Third-Party Host with No ILMI and No RFC-1577
Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 12
Figure 2.2 Configuring a Third-Party Switch with ILMI Support and No
RFC-1577 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 13
Figure 2.3 Configuring a Third-Party Switch with RFC-1577 and No ILMI
Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 14

Figure 3.1 Basic Emulated LAN Interconnections . . . . . . . . . . . . . . . . . . . . . 3 - 2
Figure 3.2 ELAN Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 5
Figure 3.3 Single Server LANE Services Model . . . . . . . . . . . . . . . . . . . . . . 3 - 9
Figure 3.4 Broadcast IP-ARP Request . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 10
Figure 3.5 IP ARP Response Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 10
ATM Switch Network Configuration Manual LOF - 1

List of Figures
Figure 3.6 Distributed LAN Emulation Model . . . . . . . . . . . . . . . . . . . . . . . 3 - 12

Figure 3.7 IP ARP Broadcast from LEC 1 to LEC 9 . . . . . . . . . . . . . . . . . . 3 - 13
Figure 3.8 Re-distributing the Broadcast across DLE Peer Servers . . . . . . 3 - 13
Figure 3.9 LE-ARP for Unknown Host Sent to Proxies (not shown) and
DLE Peer Servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 14
Figure 3.10 LE-ARP Query Answered by One DLE Peer Server and
Re-distributed by Another . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 14
Figure 3.11 LE-ARP Response Delivered and LEC 9 Contacts LEC 1 . . . . 3 - 15
Figure 3.12 ELAN with a Single Server and Multiple Switches Connecting
to Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 16
Figure 3.13 ELAN with Single Server and Remote Connection to
Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 17
Figure 3.14 ELAN with Single Server in Operation. . . . . . . . . . . . . . . . . . . . 3 - 17
Figure 3.15 Registrations on an ELAN with Multiple Servers . . . . . . . . . . . . 3 - 18
Figure 3.16 ELAN with Multiple Servers in Operation. . . . . . . . . . . . . . . . . . 3 - 19
Figure 3.17 Failure of One ELAN Server and the Recovery Process . . . . . . 3 - 19
Figure 3.18 ELAN Re-established Using the Second Server . . . . . . . . . . . . 3 - 20
Figure 3.19 Sample LECS Configuration File (Part One of Two) . . . . . . . . . 3 - 37
Figure 3.20 Sample LECS Configuration File (Part Two of Two) . . . . . . . . . 3 - 38
CHAPTER 4 MPOA
Figure 4.1 An Example of an ELAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 4
Figure 4.2 LANE without MPOA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 7
Figure 4.3 LANE with MPOA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 9
Figure 4.4 MPOA Example Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 - 11

Figure 5.1 Example of a 13-byte Switch Prefix . . . . . . . . . . . . . . . . . . . . . . . 5 - 3
Figure 5.2 Private ATM Network with 21 Switches and 34 Bidirectional
Links . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 6
Figure 5.3 Example of FT-PNNI Hierarchy Showing Lowest-Level Peer
Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 - 7
Figure 5.4 View of the Network from Switches in Peer Group A. . . . . . . . . 5 - 10

Figure 6.1 Peer Groups Summarized into a Single Address . . . . . . . . . . . . 6 - 4
Figure 6.2 PGL Hierarchy Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 6
Figure 6.3 Example of a Lop-sided Hierarchy. . . . . . . . . . . . . . . . . . . . . . . . 6 - 7
Figure 6.4 Aggregation Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 14
LOF - 2 ATM Switch Network Configuration Manual

List of Figures
Figure 6.5 Internetworking of FT-PNNI and PNNI . . . . . . . . . . . . . . . . . . . . 6 - 17

Figure 6.6 Split Switches and Gateway Switches Connecting Areas . . . . . 6 - 19
Figure 6.7 Peer Groups Connected by Border Links . . . . . . . . . . . . . . . . . . 6 - 20
Figure 6.8 Inter-domain Route Connecting Two Domains . . . . . . . . . . . . . . 6 - 21
Figure 6.9 Example Network to be Configured . . . . . . . . . . . . . . . . . . . . . . 6 - 35
Figure 6.10 LGN A.X Represents Peer Group A.X at Level 72 . . . . . . . . . . . 6 - 40
Figure 6.11 Completed Peer Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 41
Figure 7.1 UNI 4.0 Supplementary Services . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 2
Figure 7.2 DSL-to-ATM Network Connection. . . . . . . . . . . . . . . . . . . . . . . . . 7 - 8
Figure 7.3 Virtual UNI Network Configuration . . . . . . . . . . . . . . . . . . . . . . . 7 - 10
Figure 7.4 Example of a Proxy Signalling Configuration . . . . . . . . . . . . . . . 7 - 14
CHAPTER 8 Security
APPENDIX B Configuring Circuit Emulation Services
APPENDIX C Converting FT-PNNI and PNNI Networks

Figure C.1 Invalid Configuration of ASX-1000, ASX-1200, or TNX-1100 Split
between Two FT-PNNI Peer Groups . . . . . . . . . . . . . . . . . . . . . .C - 2
Figure C.2 Invalid Configuration of ASX-1000, ASX-1200, or TNX-1100 Split
between Two PNNI Peer Groups . . . . . . . . . . . . . . . . . . . . . . . . .C - 4
Figure C.3 Multiple Fabrics of an ASX-1000, ASX-1200, or TNX-1100
Incorrectly Configured as Gateways . . . . . . . . . . . . . . . . . . . . . . .C - 5
Figure C.4 A Non-Hierarchical FT-PNNI Network. . . . . . . . . . . . . . . . . . . . . .C - 7
Figure C.5 S5 Changed to a Gateway Switch . . . . . . . . . . . . . . . . . . . . . . . .C - 8
Figure C.6 S3 Changed to a Gateway Switch . . . . . . . . . . . . . . . . . . . . . . . .C - 8
Figure C.7 S4 Changed to a Gateway Switch and S5 to PNNI . . . . . . . . . . .C - 9
Figure C.8 A Completely Converted PNNI Network . . . . . . . . . . . . . . . . . . .C - 10
Figure C.9 Hierarchical FT-PNNI Network with 3 Peer Groups and a
Contiguous Backbone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .C - 11
Figure C.10 C.1 and A.1 as Gateway Switches . . . . . . . . . . . . . . . . . . . . . . .C - 13
Figure C.11 Peer Group Severed from the Rest of the FT-PNNI Area . . . . . .C - 14
Figure C.12 Peer Group C before the Conversion to PNNI . . . . . . . . . . . . . .C - 16
Figure C.13 C.1 and C.2 Not Part of Peer Group C . . . . . . . . . . . . . . . . . . . .C - 18
Figure C.14 A Completely Converted PNNI Network . . . . . . . . . . . . . . . . . . .C - 20

List of Figures
Figure C.15 C.6 as a Gateway. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 22

Figure C.16 Peer Group C Fully Migrated to PNNI . . . . . . . . . . . . . . . . . . . . C - 24
Figure C.17 Peer Group B Disconnected from Peer Group A . . . . . . . . . . . . C - 27
Figure C.18 A Migrated Backbone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 28
Figure C.19 Hierarchical FT-PNNI Network with 3 Peer Groups and a
Non-contiguous Backbone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 29
Figure C.20 Hierarchical PNNI Network after Migration . . . . . . . . . . . . . . . . C - 30
Figure C.21 Non-hierarchical PNNI Example . . . . . . . . . . . . . . . . . . . . . . . . C - 32
Figure C.22 B.X.1 Elected as PGL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 33
Figure C.23 Migrating the Links. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 34
Figure C.24 Node B.1 is De-activated and LGN B.X Represents Peer
Group B.X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 35
Figure C.25 Completed Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C - 36

Figure D.1 Buffer Sizes Configured Using the Hardware Netmod Fram
Modify Command. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-4
Figure D.2 CLP0PPD Automatically Calculated . . . . . . . . . . . . . . . . . . . . . . D-5
Figure D.3 Calculating CLP1EPD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-6
Figure D.4 Calculating CLP0EPD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-7
Figure D.5 Calculating CLP1PPD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D-8
APPENDIX E RMON

Figure F.1 Multiple Physical Links Aggregated into One Logical Link. . . . . . F - 2

Figure G.1 Linear APS Failover to the Protection Link upon Detection of
SONET Line Failure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 2
Figure G.2 Port Redundancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 4
Figure G.3 Port Card Redundancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 4
Figure G.4 1+1 Data Path Protection (Half of Switch Is Shown) . . . . . . . . . . G - 6
Figure G.5 ASX-4000 20 Gbps Fully Redundant Configuration . . . . . . . . . . G - 9
Figure G.6 ASX-4000 10 Gbps Redundant and 20 Gbps Non-redundant
Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 10
Figure G.7 ASX-4000 10 Gbps Redundant Configuration. . . . . . . . . . . . . . G - 11

List of Figures
Figure G.8 ASX-4000 Line Level and Port Card Redundancy without Fabric
Redundancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 12
Figure G.9 ASX-4000 Fabric Redundancy without Line Level and Port Card
Redundancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G - 13

List of Figures

List of Tables

Table 1.1 Summary of Traffic Contract Variables and Policing
Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 40

Table 3.1 LECS Configuration File Parameters . . . . . . . . . . . . . . . . . . . . . 3 - 26

Table 6.1 Example of Peer Group Addressing . . . . . . . . . . . . . . . . . . . . . . . 6 - 5
Table 6.2 Example Enterprise NSAP Prefix . . . . . . . . . . . . . . . . . . . . . . . . 6 - 33
Table 6.3 Example Service Provider NSAP Prefix . . . . . . . . . . . . . . . . . . . 6 - 34
Table 7.1 Action Taken Based on Both Switches’ Signalling Channel
Configurations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 26
Table 7.2 Action Taken Based on the Peer’s Supported MIB Variable . . . . 7 - 27
Table 7.3 Valid Type and Version Combinations. . . . . . . . . . . . . . . . . . . . . 7 - 28
Table 7.4 Invalid Type and Version Combinations . . . . . . . . . . . . . . . . . . . 7 - 30
Table 7.5 Valid Scope and Mode Combinations. . . . . . . . . . . . . . . . . . . . . 7 - 31
Table 7.6 Invalid Scope and Mode Combinations . . . . . . . . . . . . . . . . . . . 7 - 33
Table 7.7 UNI 3.1 Allowable Combination of Traffic Parameters in
ForeThought 5.2.x and Greater . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 37
CHAPTER 8 Security
Table 8.1 Allowable Login Method Combinations. . . . . . . . . . . . . . . . . . . . . 8 - 2
Table 8.2 Possible Login Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 3

Table A.1 SNMP Traps Supported on the FORE Switches . . . . . . . . . . . . .A - 2
Table A.2 Message Type Encodings for Trap 2003. . . . . . . . . . . . . . . . . . .A - 21
Table A.3 Error Codes for Trap 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .A - 22

Table D.1 Buffer Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .D - 3
ATM Switch Network Configuration Manual LOT - 1

List of Tables
APPENDIX E RMON
Table E.1 Common Operations Supported by SNMP . . . . . . . . . . . . . . . . . E - 3
LOT - 2 ATM Switch Network Configuration Manual

Preface
Preface
This manual provides the technical information needed to configure the ForeRunner ® family of
ATM Switches, ForeRunnerLE ® Switches, TNX™ ATM Switches, ESX™-3000 Switch, the
ForeRunner network modules, and the accompanying ForeThought™ software. This document
also provides general ATM information and general product information. This document was
created for users with various levels of experience. If you have any questions or problems,
please contact the Marconi Technical Assistance Center (TAC).
Chapter Summaries
Chapter 1 - Configuring PVCs - Describes how to create PVCs on a switch through the ATM
Management Interface (AMI).
Chapter 2 - Configuring Classical IP - Describes how to design, configure, and maintain a
Classical IP ATM network.
Chapter 3 - Configuring an Emulated LAN - Provides an overview of LAN Emulation and
gives an example of how to configure an Emulated LAN.
Chapter 4 - MPOA - Contains an overview of LAN Emulation (LANE) and Multi-Protocol
Over ATM (MPOA).
Chapter 5 - ForeThought PNNI - Provides an overview of ForeThought PNNI and shows how
this scalable routing and signalling protocol can be used to simplify large network topologies.
Chapter 6 - ATM Forum PNNI - Provides an overview of ATM Forum PNNI and shows how
Preface
this scalable routing and signalling protocol can be used to simplify large network topologies.
Chapter 7 - Signalling - Describes signalling protocol information.
Chapter 8 - Security - Describes the various forms of security that can be used on the switch.
Chapter 9 - Configuring Timing - Describes how to set up timing on a switch.
Appendix A - Configuring SNMP - Describes the remote SNMP configuration of a switch.
Appendix B - Configuring Circuit Emulation Services - Contains information for configur-
ing Circuit Emulation Services (CES) network modules.
ATM Switch Network Configuration Manual i

Preface
Appendix C - Converting FT-PNNI and PNNI Networks - Discusses the conversion of both
non-hierarchical and hierarchical FT-PNNI networks to ATM Forum PNNI networks. Also
describes the conversion of a non-hierarchical PNNI network to a lop-sided hierarchical PNNI
network.
Appendix D - Configuring FramePlus Modules - Contains information for configuring
FramePlus™ network modules.
Appendix E - RMON - Contains information for configuring RMON.
Appendix F - Configuring IMA Modules - Contains information for configuring IMA net-
work modules.
Appendix G - ASX-4000 Redundancy - Provides an overview of SONET automatic protec-
tion switching (APS) and the ASX-4000 fabric and portcard redundancy features.
Related Manuals
References are made in this manual to the following manuals:
ATM Management Interface (AMI) Manual, Part 1 - Describes the root, connections, ethernet,
and hardware level AMI commands and menus.
ATM Management Interface (AMI) Manual, Part 2 - Describes the interfaces, redundancy, rout-
ing, and security level AMI commands and menus.
ATM Management Interface (AMI) Manual, Part 3 - Describes the services, signalling, and sys-
tem level AMI commands and menus.
ATM Switch Diagnostics and Troubleshooting Manual - Describes the debug level AMI com-
mands and menus. Also, describes error messages, loopbacks, SCP diagnostics, and ATM
Forum PNNI debugging information.
These manuals can be found on the CD and can be read and printed using Acrobat Reader
which is also included on the CD. If Acrobat Reader is installed locally, run Acrobat and open
the manual from the /DOCS directory of the CD. If Acrobat Reader is not installed locally, run
the Acrobat installer to load Acrobat Reader on your machine. Then run the ACROREAD.EXE
file in the /DOCS directory of the CD.
ii ATM Switch Network Configuration Manual

Preface
Technical Support
In the U.S.A., customers can reach Marconi Technical Assistance Center (TAC) using any one
of the following methods:
1. Select the “Support” link from Marconi’s World Wide Web page:
http://www.marconi.com/
2. Send questions, via e-mail, to:

[email protected]
3. Telephone questions to “support” at:

800-671-FORE (3673) or 724-742-6999
4. FAX questions to “support” at:

724-742-7900
Technical support for customers outside the United States should be handled through the local
distributor or via telephone at the following number:
+1 724-742-6999
No matter which method is used to reach Marconi Support, customers should be ready to pro-
vide the following:
• A support contract ID number
Preface
• The serial number of each product in question
• All relevant information describing the problem or question
ATM Switch Network Configuration Manual iii

Preface
Typographical Styles
Throughout this manual, all specific commands meant to be entered by the user appear on a
separate line in bold typeface. In addition, use of the Enter or Return key is represented as
<ENTER>. The following example demonstrates this convention:
cd /usr <ENTER>
File names that appear within the text of this manual are represented in the following style:
“...the fore_install program installs this distribution.”
Command names that appear within the text of this manual are represented in the following
style: “...using the flush-cache command clears the bridge cache.”
Subsystem names that appear within the text of this manual are represented in the following
style: “...to access the bridge subsystem...”
Parameter names that appear within the text of this manual are represented in the following
style: “...using <seg-list> allows you to specify the segments for which you want to display
the specified bridge statistics.”
Any messages that appear on the screen during software installation and network interface
administration are shown in Courier font to distinguish them from the rest of the text as fol-
lows:
.... Are all four conditions true?
iv ATM Switch Network Configuration Manual

Preface
Important Information Indicators

To call your attention to safety and otherwise important information that must be reviewed to
ensure correct and complete installation, as well as to avoid damage to the Marconi product or
to your system, Marconi utilizes the following WARNING/CAUTION/NOTE indicators.
WARNING statements contain information that is critical to the safety of the operator and/or
the system. Do not proceed beyond a WARNING statement until the indicated conditions are
fully understood or met. This information could prevent serious injury to the operator, dam-
age to the Marconi product, the system, or currently loaded software, and is indicated as fol-
lows:
WARNING! Hazardous voltages are present. To reduce the

risk of electrical shock and danger to personal
health, follow the instructions carefully.
CAUTION statements contain information that is important for proper installation/opera-

tion. Compliance with CAUTION statements can prevent possible equipment damage and/
or loss of data and are indicated as follows:
CAUTION You risk damaging your equipment and/or

software if you do not follow these instructions.
Preface
NOTE statements contain information that has been found important enough to be called to
the special attention of the operator and is set off from the text as follows:
If you change the value of the LECS control

NOTE parameters while the LECS process is running,
the new values do not take effect until the LECS
process is stopped, and then restarted.
ATM Switch Network Configuration Manual v

Preface
Laser Warning
Every Marconi network module having a single mode fiber optic interface contains a Class 1
laser.
Class 1 Laser Product:

This product conforms to
applicable requirements of
21 CFR 1040 at the date of
manufacture.
Class 1 lasers are defined as products which do not permit human access to laser radiation in
excess of the accessible limits of Class 1 for applicable wavelengths and durations. These
lasers are safe under reasonably foreseeable conditions of operation.
WARNING! Do not stare into the beam or view the beam

with optical instruments.
vi ATM Switch Network Configuration Manual

Preface
Safety Precautions
For your protection, observe the following safety precautions when setting up equipment:
• Follow all warnings and instructions marked on the equipment.
• Ensure that the voltage and frequency of your power source matches the voltage
and frequency inscribed on the equipment’s electrical rating label.
• Never push objects of any kind through openings in the equipment. Dangerous
voltages may be present. Conductive foreign objects could produce a short circuit
that could cause fire, electric shock, or damage to your equipment.
Modifications to Equipment
Do not make mechanical or electrical modifications to the equipment. Marconi is not responsi-
ble for regulatory compliance of a modified Marconi product.
Placement of a Marconi Product
CAUTION To ensure reliable operation of your Marconi

product and to protect it from overheating,
openings in the equipment must not be blocked
or covered. A Marconi product should never be
placed near a radiator or heat register.
Preface
ATM Switch Network Configuration Manual vii

Preface
Power Cord Connection
WARNING! Marconi products are designed to work with

single-phase power systems having a grounded
neutral conductor. To reduce the risk of electrical
shock, do not plug Marconi products into any
other type of power system. Contact your
facilities manager or a qualified electrician if you
are not sure what type of power is supplied to
your building.
WARNING! Your Marconi product is shipped with a

grounding type (3-wire) power cord. To reduce
the risk of electric shock, always plug the cord
into a grounded power outlet.
viii ATM Switch Network Configuration Manual

To establish a permanent communication link between two sites, it is necessary to establish

permanent virtual circuits (PVCs) at each switch along the communications path. This chapter
discusses the creation of PVCs through the ATM Management Interface (AMI), a command-
line user interface to ForeRunner ATM switches, ForeRunnerLE ATM switches, TNX ATM
switches, and ESX-3000 ATM switches.
Configuring PVCs
1.1 General Concepts
Each ATM cell contains a virtual path identifier (VPI) and a virtual channel identifier (VCI) as
part of its five-byte ATM header. The VPI and VCI are used to route the cell through the ATM
network. When a switch fabric receives a cell, it examines the ATM header to determine the
correct output port, VPI, and VCI for the cell. For example, an ATM switch fabric can be con-
figured such that any cell received on port A1 with VPI|VCI = 0|32 is switched to port B2
with VPI|VCI = 0|35. The translation from input port, VPI, and VCI to output port, VPI, and
VCI is achieved via a mapping table in the switch fabric’s memory.
The VCI value of cells does not change as the cell is switched through the ATM network via a
virtual path. In a single switch environment, a cell’s VPI and VCI are translated only once, but
in a multiple switch environment a cell’s VPI and VCI are translated many times. It is impor-
tant to remember that a cell’s VPI and VCI are of local significance only (i.e., link-by-link). It is
also important to note that virtual connections are unidirectional; that is, they are valid in one
direction only. The VPI and VCI may change as the cell is switched through the network.
ATM Switch Network Configuration Manual 1-1

Configuring PVCs
VPI: 1, VCI: 37
cell
FORE ATM FORE ATM
Switch B Switch C
port C1 port B4 port D1 port B1
port A4 VPI: 2, VCI: 33 cell port A1

cell VPI: 1, VCI: 35
FORE ATM FORE ATM

Switch A Switch D
cell VPI: 0, VCI: 32 VPI: 0, VCI: 36 cell

port C1 port B3
Workstation P Workstation Q
Figure 1.1 - The Cell
The mappings in an ATM network used to route cells from a source to a destination are gener-
ally referred to as virtual channels and virtual paths. The following sections is to explain how
to create the necessary mappings to establish these virtual paths and virtual channels in a net-
work of FORE ATM switches.
1-2 ATM Switch Network Configuration Manual

Configuring PVCs
1.2 Virtual Paths

Virtual paths, which are carried within a physical transit medium (e.g., DS1, E1, DS3, E3,
OC3c, or OC12c link), are used to establish connections between two nodes in an ATM net-
work. Many virtual paths can be transmitted within a single physical link. Two types of vir-
tual paths exist: virtual path connections (VPCs), also known as through paths, and
originating/terminating paths, also known as virtual path terminators (VPTs). VPCs allow
virtual paths to be cross-connected at a switch node while VPTs allow virtual channels (VCCs)
to be cross-connected or switched at a switch node.
Configuring PVCs
Virtual
Path
Virtual
Channels
Medium
Figure 1.2 - Virtual Channels in a Virtual Path
A single virtual path can be used to route many virtual channels through the ATM network.
Because a virtual path simply routes virtual channels through the network, a cell is guaran-
teed to have the same VCI when it exits the virtual path as it had when it entered the virtual
path.
FORE ATM Switch ATM FORE ATM Switch

Network
cell cell
VPI: X, VCI: Y VPI: Z, VCI: Y
Figure 1.3 - An Example of a Virtual Path

Configuring PVCs
The VCI value of cells does not change as the cell is switched through the ATM network via a
virtual path. Each virtual path must originate at a switch fabric, pass through zero or more
switch fabrics and terminate at another switch fabric. The origination and termination points
are referred to as originating and terminating paths. Virtual paths are switched through switch
fabrics via through paths. Virtual paths are made up of an originating path, zero or more
through paths, and a terminating path.
Originating Path Through Path Terminating Path
ForeRunner
ATM Switch
FORE FORE
ATM Switch ATM Switch
Virtual Path
Figure 1.4 - Composition of a Virtual Path
1.2.1 Through Paths

Through paths route an entire virtual path through an ATM switch fabric. When a cell is
received by a switch fabric on a through path, the VPI is examined to determine the output
port and VPI. The VCI component of the ATM header remains unchanged and can have any
value. So, all of the channels within the through path are switched correctly without altering
the VCI value of cells on these channels.
Four parameters are needed to define a through path on a FORE switch fabric: input ATM
interface, input VPI, output ATM interface, and output VPI.
The VCI value remains unchanged when cells are switched via a through path. For example,
the through path A4|10 -> B4|20 maps cells received on port A4 with VPI: 10 and any VCI to
port B4 with VPI: 20 and the same VCI.

Configuring PVCs
switch fabric
A4 B4
cell cell
VPI: 10 VPI: 20
VCI: X Through Path VCI: X
A4|10 -> B4|20
Configuring PVCs
Figure 1.5 - An Example of a Through Path
By definition, through paths only switch cells in one direction; they are unidirectional. For
example, switch fabric X is configured with the through path B1|20 -> C1|20. If a cell is
received on port C1 with VPI: 20, it is not transmitted on port B1 with a new VPI: 20. In order
for this to happen, the through path C1|20 -> B1|20 must exist as well. Since through paths
are unidirectional, two through paths are necessary for bidirectional communication.
B1|20 -> C1|20

B1 C1
C1|20 -> B1|20
Figure 1.6 - Through Paths are Unidirectional

Configuring PVCs
1.2.2 Originating and Terminating Paths

As previously noted, originating and terminating paths (also called virtual path terminators)
are points at which a virtual path originates and terminates. For example, if a virtual path
exists from switch fabric A to switch fabric B, then there must be an originating path on switch
fabric A and a terminating path on switch fabric B.
An originating path is defined by two parameters: output VPI and output port. Similarly, a
terminating path is defined by the parameters: input VPI and input port. Because originating
and terminating paths do not define the way cells are switched through an ATM switch fabric,
virtual channels must exist to switch cells from a terminating path to an originating path. (See
the section about virtual channels for more information.) Originating and terminating paths
are the endpoints of virtual paths and are used primarily for bandwidth allocation.
The bandwidth allocated to originating and terminating paths is used to control the amount of
virtual channel (VCC) bandwidth entering or leaving a virtual path. The total guaranteed
bandwidth used by virtual channels on an originating path or a terminating path cannot
exceed the amount of bandwidth allocated to that path. For example, as illustrated in
Figure 1.7, if each of the four virtual channels shown is using 10 Mbps of bandwidth, then the
originating and terminating paths must have at least 40 Mbps of bandwidth allocated.
UBR traffic bandwidth, which is a “best effort”

NOTE service class, is not limited by the VP’s allocated
bandwidth since its bandwidth is not
guaranteed. Actual UBR VCC traffic transmitted
within a VP may exceed the VP’s allocated
bandwidth.
Originating Path Terminating Path
FORE FORE
ATM Switch ATM Switch
Virtual Channels
Figure 1.7 - Using Originating and Terminating Paths for Bandwidth Allocation

Configuring PVCs
1.3 Listing Virtual Paths
1.3.1 Listing Through Paths

By logging in to AMI (see Chapter 1 of Part 1 of the AMI Configuration Commands Reference
Manual for information about logging into AMI), it is possible to display either all of the exist-
ing through paths on an individual switch fabric or all of the existing through paths on a spec-
ified port. To list all of the existing through paths on an individual switch fabric, enter the
following:
Configuring PVCs
connections path through show
Input Output
AtmIf VPI AtmIf VPI AllocBW Serv Protocol Name
Categy
1A1 1 1A1 2 0.0K UBR pvc N/A
1A1 4 1A1 5 0.0K UBR pvc N/A
1A1 5 1A1 6 0.0K UBR pvc N/A
1A1 7 1A1 6 0.0K UBR pvc N/A
The fields in this display are defined as follows:
Field Description
AllocBW The amount of bandwidth reserved on this path, in Kbps.
Input AtmIf The incoming port number of the through path.
Input VPI The incoming virtual path number.
Output AtmIf The outgoing port number of the through path.
Output VPI The outgoing virtual path number.
AllocBW The amount of bandwidth reserved on this path, in Kbps.
Serv Categy The traffic type on this virtual path.
Protocol The type of protocol running on this path.
Name The user-assigned name which helps to identify this through path uniquely.

Configuring PVCs
To list advanced options about all of the existing virtual (through) paths, enter the following
parameters:
connections path through show -advanced
Input Output
AtmIf VPI AtmIf VPI LoopVPI UPC ConType
1A1 1 1A1 2 0 N/A
1A1 4 1A1 5 0 N/A
1A1 5 1A1 6 0 N/A
1A1 7 1A1 6 0 N/A
The fields in the advanced display are defined as follows:
Field Description
Input AtmIf The incoming port number of the through path.
Output AtmIf The outgoing port number of the through path.
LoopVPI Indicates whether or not traffic shaping has been enabled for this path.
UPC The integer index that refers to a specific traffic contract assigned to this through path.
UPC contracts can be displayed using connections upc show.
ConType The connection type for the endpoints of this path with respect to a particular network.
Orig (originating) means that the ingress/egress endpoint of the path is connected to the
source node which is outside the network, tran (transit) means that the ingress/egress
endpoint of the path is connected to a node within the network, and term (terminating)
means that the ingress/egress endpoint of the path is connected to the destination node
which is outside the network.

Configuring PVCs
1.3.2 Listing Originating and Terminating Paths

By logging in to AMI, it is possible to display either all of the existing originating and termi-
nating paths on an individual switch fabric or on a specified port. To list all of the originating
and terminating paths on an individual switch fabric, enter the following parameters:
connections path term show
AtmIf VPI Type ResBW CurBW MinVCI MaxVCI VCs Protocol

4A1 0 term N/A 356.2K 1 511 7 pvc
4A1 0 orig N/A 355.7K 1 511 7 pvc
Configuring PVCs
4A2 0 term N/A 0.8K 1 511 6 pvc
4A2 0 orig N/A 0.0K 1 511 6 pvc
4A3 0 term N/A 356.2K 1 511 7 pvc
4A3 0 orig N/A 355.7K 1 511 7 pvc
4A4 0 term N/A 0.8K 1 511 6 pvc
4A4 0 orig N/A 0.0K 1 511 6 pvc
Press q to quit, any other key to continue... q
Field Description
AtmIf The incoming port number of the vpt.
VPI The incoming or outgoing virtual path number.
Type Shows orig if it is an originating path. Shows term if it is a terminating path.
ResBW The maximum amount of bandwidth, in Kbps, that is reserved for the virtual channels
using this vpt. A value of N/A indicates that this path is an elastic path. Elastic paths allo-
cate and de-allocate bandwidth for their channels from the link.
CurBW The amount of bandwidth, in Kbps, that is being used by the virtual channels using this
vpt.
MinVCI The bottom number for the range of VCIs that are reserved for VCCs on this virtual path
terminator.
MaxVCI The top number for the range of VCIs that are reserved for VCCs on this virtual path ter-
minator. Also, this is the maximum VCI number to be supported on the control port.
VCs The number of virtual channels that are currently using this vpt.
Protocol The type of protocol running on this channel.

Configuring PVCs
To list the advanced options about the existing virtual path terminators, enter the following:
connections path term show -advanced
AtmIf VPI Type ShapeVPI LoopVPI VBROB BuffOB Shaping

4A1 0 term N/A N/A N/A N/A N/A
4A1 0 orig port port flat
Field Description
AtmIf The incoming port number of the vpt.
VPI The incoming or outgoing virtual path number.
Type Shows orig if it is an originating path. Shows term if it is a terminating path.
ShapeVPI Shows the output port on which traffic shaping has been enabled for this originating vpt.
This only applies to Series C network modules. N/A means that shaping has not been con-
figured on this port.
LoopVPI Shows the looping port that has been configured to loop traffic on that port for shaping on
a Series D port. N/A means that looping has not been configured on this port.
VBROB The bandwidth overbooking level assigned to this vpt, specified as a percentage. The
default is 100, which means that no overbooking has been defined. Values less than 100
cause underbooking. Values greater than 100 denote overbooking. port means this is an
elastic path. Since elastic paths derive their overbooking factors from their parent ports,
use hardware port show or interfaces atmif show to show the overbooking value.
BuffOB The buffer overbooking level assigned to this vpt, specified as a percentage. The default is
100, which means that no overbooking has been defined. Values less than 100 cause
underbooking. Values greater than 100 denote overbooking. port means this is an elastic
path. Since elastic paths derive their overbooking factors from their parent ports, use
hardware port show or interfaces atmif show to display the overbooking value.
Shaping flat means there is no shaping. All elastic VPTs are flat.
shaped means the VCCs can be serviced by any shaping mode.
shaped-roundrobin means the VCCs on this VPT are only serviced by the roundrobin
mode.
This option only applies to the Series 2 port cards on an ASX-4000 switch.
For more information about setting overbooking

NOTE parameters (VBROB and BuffOB), see Section
1.5.2.
1 - 10 ATM Switch Network Configuration Manual

Configuring PVCs
1.4 Virtual Channels

Virtual channels “ride” inside of virtual paths. The combination of the two specifies a virtual
connection. On a switch fabric, each virtual channel switches cells with a specific VPI and VCI
received on a specific port to another port with a new VPI and a new VCI. Unlike through
paths, which carry one or more VCCs, virtual channels describe a single virtual connection
between two endpoints.
Configuring PVCs
FORE
cell ATM Switch cell
VPI: X, VCI: Y VPI: A, VCI: B
Figure 1.8 - An Example of a Virtual Channel
Six parameters are needed to define a virtual channel: input port, input VPI, input VCI, out-
put port, output VPI, and output VCI. Virtual channels switch cells using both the VPI and
VCI values. Both the VPI and VCI values may change when a cell is switched via a virtual
channel. For example, the virtual channel C2|1|20 -> D2|9|25 switches cells received on port
C2 with VPI: 1 and VCI: 20 such that they are transmitted out port D2 with VPI: 9 and VCI: 25.
switch fabric
C2 D2
cell cell
VPI: 1 VPI: 9
VCI: 20 VCI: 25
Virtual Channel
C2|1|20 -> D2|9|25
Figure 1.9 - Example of a Virtual Channel
ATM Switch Network Configuration Manual 1 - 11

Configuring PVCs
To establish two-way communications between two ports on a switch fabric, two virtual chan-
nels are necessary because virtual channels are unidirectional. For example, switch fabric A is
configured with the virtual channel C3|7|12 -> D1|8|2. If a cell is received on port D1 with
VPI: 8 and VCI: 2, it is not transmitted out port C3 with VPI: 7 and VCI: 12. An additional
channel, namely D1|8|2 -> C3|7|12, would have to exist.
C3|7|12 -> D1|8|2

C3 D1
D1|8|2 -> C3|7|12
Figure 1.10 - Virtual Channels are Unidirectional
Before a virtual channel can be created, the corresponding terminating and originating paths
must exist. For example, before the channels shown on the switch fabric in Figure 1.11 can be
created, the terminating path C3|3 must exist.
Switch or
FORE Host B
ATM Switch
Term. Orig. Switch or
Host C
C3|3|45 -> A3|9|100
Switch or C3|3|50 -> C2|3|98 Switch or
Host A C3|3|80 -> A1|7|88 Host D
C3|3|123 -> A1|3|123
Switch or
Host E
Figure 1.11 - Virtual Channels Created on Terminating Path C3|3
Similarly, before the virtual channels shown in Figure 1.12 can be created, the originating path
C2|2 must exist.

Configuring PVCs
Switch or
Host A FORE
ATM Switch
Switch or Term. Orig.
Host B
A2|7|120 -> C2|2|120
Switch or C2|3|67 -> C2|2|37 Switch or
Host E
Host C C3|11|50 -> C2|2|102
B1|2|99 -> C2|2|99
Switch or
Configuring PVCs
Host D
Figure 1.12 - Virtual Channels Created on Originating Path C2|2
Furthermore, in these examples, the terminating path C3|3 and originating path C2|2 must
have enough bandwidth allocated to support the total bandwidth used by the virtual channels
(see Figure 1.7).
1.4.1 Smart Permanent Virtual Circuits

Smart Permanent Virtual Circuits (SPVCs) are connections that go across multiple switch fab-
rics. An SPVC looks like a PVC at the local and remote endpoints with an SVC (Switched Vir-
tual Circuit) in the middle. SVCs are channels established on demand by network signalling.
Similar to a dialed telephone call, SVCs transport information between two locations and last
only for the duration of the transfer.
SPVCs are more robust than PVCs. If a link carrying a PVC goes down, then the PVC goes
down. If a link carrying a SPVC goes down and there is an alternate route, then the end switch
fabrics of the SPVC automatically reroute the SPVC around the failed link.
As shown in Figure 1.13, interswitch links exist between the FORE switches. The endpoints
exist at switch A and switch E. If a link goes down between switch A and switch B, an SPVC
can reroute the cell (via an SVC) through switch C.

Configuring PVCs
FORE Switch FORE Switch

B D
VPI: 2, VCI: 33 cell

cell VPI: 1, VCI: 35
port A4 port A1
FORE Switch
C
FORE Switch FORE Switch

A E
SVCs
Interswitch Links
Figure 1.13 - The Path of a Cell Via SPVCs

Configuring PVCs
1.4.2 Listing Virtual Channels

By logging in to AMI, you can display either all of the existing virtual channels on an individ-
ual switch fabric or on a specified port. To list all of the virtual channels on an individual
switch fabric, enter the following parameters:
connections channel show
Input Output
AtmIf VPI VCI AtmIf VPI VCI ServCat Protocol Name
1A1 0 5 1CTL 0 37 nrtVBR fsig N/A
Configuring PVCs
1A1 0 14 1CTL 0 36 UBR spans N/A
1A1 0 15 1CTL 0 35 nrtVBR spans N/A
Field Description
Input AtmIf The incoming port number of the virtual channel.
Input VCI The incoming virtual channel number.
Output AtmIf The outgoing port number of the virtual channel.
Output VCI The outgoing virtual channel number.
ServCat The traffic type on this virtual channel.
Protocol Indicates what type of signalling protocol is running on this channel. Can be spans, pvc,
fsig, spvc, or rcc. rcc is the routing control channel (0, 18) on PNNI links over which
PNNI exchanges routing information. fsig stands for ATM Forum signalling.
Name The unique, user-assigned name for this PVC. If no name is assigned, shows N/A.

Configuring PVCs
To list advanced information about all of the existing permanent virtual channels on a switch
board, enter the following:
connections channel show -advanced
Input Output
AtmIf VPI VCI AtmIf VPI VCI UPC Protocol ConType
1A1 0 5 1CTL 0 37 N/A fsig N/A
1A1 0 14 1CTL 0 36 N/A spans N/A
1A1 0 16 1CTL 0 38 N/A fsig N/A
1A2 0 16 1CTL 0 42 N/A fsig N/A
1A3 0 16 1CTL 0 46 N/A fsig N/A
Field Description
Input AtmIf The incoming port number of the virtual channel.
Input VCI The incoming virtual channel number.
Output AtmIf The outgoing port number of the virtual channel.
Output VCI The outgoing virtual channel number.
UPC The integer index that refers to the specific UPC traffic contract assigned to this VCI.
Protocol Indicates what type of signalling protocol is running on this channel. Can be spans, pvc,
fsig, spvc, or rcc. rcc is the routing control channel (0, 18) on PNNI links over which
PNNI exchanges routing information. fsig stands for ATM Forum signalling.
ConType The connection type for the endpoints of this PVC with respect to a particular network.
Orig (originating) means that the ingress/egress endpoint of the channel is connected to
the source node which is outside the network, tran (transit) means that the ingress/
egress endpoint of the PVC is connected to a node within the network, and term (termi-
nating) means that the ingress/egress endpoint of the PVC is connected to the destination
node which is outside the network.

Configuring PVCs
1.5 Creating PVCs and SPVCs

FORE’s ATM network modules provide ATM transmission connectivity, while its intelligent
network modules, such as the Circuit Emulation Services (CES) network module, provide
adaptation for ports carrying one transmission format (e.g., TDM) to ATM cells. This section
describes how to create the following from one ATM port to another:
• a permanent virtual path (through path)
• a permanent virtual path terminator (originating or terminating path)
• a permanent virtual channel through the network
Configuring PVCs
• a smart permanent virtual channel through the network
This section assumes that the physical port parameters of the switches have already been con-
figured and that the ATM network module traffic models have been set appropriately. For
more information about these configurations, see Part 1 of the AMI Configuration Commands
Reference Manual.
When these paths and channels are created, a

NOTE command is entered automatically into the
current configuration database (CDB), meaning
that these paths and channels are created every
time the switch control processor (SCP) is
restarted. The CDB should be backed up
frequently.

Configuring PVCs
1.5.1 Creating a Through Path

To create a new through path, log in to AMI and enter the following parameters:
myswitch:connections path through-> new
Usage:
[-iatmif] <AtmIf> Input Port
[-ivpi] <integer> Input VPI
[-oatmif] <AtmIf> Output Port
[-ovpi] <integer> Output VPI
[[-upc] <UPC Index>] UPC Index (default: 0)
[[-loopvpi] <integer>] Loop VPI
[[-name] <text>] Name
[[-ctype] <connection_type>] Connection Type
See Part 1 of the AMI Configuration Commands Reference Manual for descriptions of the
usage parameters.
Terminating and originating paths cannot be

NOTE created across the intra-fabric ports on an
ASX-1000 switch, ASX-1200 switch, or TNX-1100
switch; only through paths can be created across
the intra-fabric ports as shown in the third
example.
The following is an example of how to create a virtual path which specifies a name:
myswitch:connections path through-> new -iatmif 3b1 -ivpi 75 -oatmif 3b5 -ovpi
75 -name customer_b
The following is an example of how to create a virtual path which specifies a name and a con-
nection type:
myswitch:connections path through-> new -iatmif 3b6 -ivpi 62 -oatmif 3b2 -ovpi
62 -name customer_c -ctype tran

Configuring PVCs
The following is an example of how to create a virtual path on an ASX-1000 switch, ASX-1200
switch, or TNX-1100 switch. To create a through path going in port 2A1, VPI 1 on the switch
board installed in slot 2 and going out port 4B1, VPI 1 on the switch board installed in slot 4,
enter the following:
myswitch:connections path through-> new -iatmif 2a1 -ivpi 1 -oatmif 2e4 -ovpi 1
myswitch:connections path through-> new -iatmif 2e4 -ivpi 1 -oatmif 2a1 -ovpi 1
myswitch:connections path through-> new -iatmif 4b1 -ivpi 1 -oatmif 4e2 -ovpi 1
myswitch:connections path through-> new -iatmif 4e2 -ivpi 1 -oatmif 4b1 -ovpi 1
Configuring PVCs
In the first line in the first pair, notice that the output port is 2E4. This is the intra-fabric port.
The 2 means the connection is coming out of the switch board in slot 2 through the intra-fabric
port. The E represents the intra-fabric port. The 4 means the connection is destined for switch
board in slot 4. 2E4 then becomes the input port in the second line.
In the first line in the second pair, notice that the output port is 4E2. This is the intra-fabric
port. The 4 means the connection is coming out of the switch board in slot 4 through the intra-
fabric port. The E represents the intra-fabric port. The 2 means the connection is destined for
switch board in slot 2. 4E2 then becomes the input port in the second line.
At the same time, a command is entered automatically into the current configuration data-
base, which means that this virtual path is created every time the SCP is restarted.

Configuring PVCs
1.5.2 Creating an Originating or Terminating Path

To create an originating or terminating path (virtual path terminator), log in to AMI and enter
the following parameters:
myswitch:connections path term-> new

Usage:
[-atmif] <AtmIf> AtmIf
[-vpi] <integer> VPI
[[-type] (orig|term)] Direction
[[-allocbw] <bps>] Allocated Bandwidth
[[-minvci] <integer>] Min VCI (default: 1)
[[-maxvci] <integer>] Max VCI (default: 511)
[[-trafficsvpi] <integer>] Traffic Shape VPI
[[-loopvpi] <integer>] Loop VPI
[[-vbrob] <percent>] VBR Overbooking
[[-vbrbob] <percent>] VBR Buffer Overbooking
[[-shaping] (flat|shaped|shaped-roundrobin)] Originating Path Shaping
(default: flat)
[[-cbr] (none|default|qosext_index)] CBR
[[-rtvbr] (none|default|qosext_index)] RtVBR
[[-nrtvbr] (none|default|qosext_index)] NrtVBR
[[-abr] (none|default|qosext_index)] ABR
[[-ubr] (none|default|qosext_index)] UBR
The qosext_index must exist (under

NOTE connections qosext show) before it can be
applied to the originating/terminating path.
usage parameters.

Configuring PVCs
Bandwidth and/or buffer overbooking are used

NOTE when the number of VBR VCCs configured
across an originating or terminating VPT exceeds
the guaranteed capacity of the VP. While
overbooking removes service guarantees (in the
case when all users transmit at full contract rate
simultaneously), a network can be engineered in
a more cost-effective manner if it is assumed that
all VCC sources do not transmit simultaneously.
Guideline: An initial buffer overbooking value
Configuring PVCs
of 500-700% is recommended until live traffic
patterns and VPC utilization can be measured. If
bandwidth is still under-utilized, a higher
VBROB setting can be used.
The following is an example of how to create a terminating path:
myswitch:connections path term-> new -atmif 3b3 -vpi 99 -type term

Would you like to create the originating side also [y]? y
The following is an example of how to create a originating path:
myswitch:connections path term-> new -atmif 3b4 -vpi 88 -type orig

Would you like to create the terminating side also [y]? y
The following is an example of how to delete a terminating path:
myswitch:connections path term-> del -atmif 3b4 -vpi 88 -type term

Would you like to delete the originating side also [y]? y
The following is an example of how to delete an originating path:
myswitch:connections path term-> del -atmif 3b3 -vpi 99 -type orig

Would you like to delete the terminating side also [y]? y
If you do not specify term or orig, the switch automatically deletes both sides of the path:
myswitch:connections path term-> del -atmif 3b4 -vpi 88
Before deleting a virtual path, you must first

NOTE delete all VCCs which use that path.

Configuring PVCs
1.5.2.1 Shaping Multiple Originating Paths on a Single Port

This feature allows you to shape several originating paths to be output on a single port. This
feature is useful if you have several remote sites interconnected by a PVP mesh. If you only
need to shape one originating path, you can simply use the hardware port traffic d
modify -ratelimit command on a Series D network module.
This feature requires the use of two ports. One port, which can be on a Series C, Series LC, or
Series D network module, is used for originating and terminating path(s). This port is put in
either physical or diagnostic loopback and is configured with the new option -loopvpi
under connections path term new. The other port provides the actual output on a shaped
through path, so this port must be on a Series D network module.
Switch B
Switch A
VP 3
WAN PVP Mesh
1A1
shaping port
VP 2
switch fabric
VP 0
VP 2
Switch C
1A2
VP 1 VP 3
looping port (orig. and term. paths VP 0 and VP 1)
Figure 1.14 - PVPs Looped through Port 1A2 and Output on Port 1A1 to WAN
As shown in the example in Figure 1.14, the traffic that is going out originating path 0 on port
1A2 gets looped so that it is output to the WAN on through path 2 on shaping port 1A1. Simi-
larly, the traffic that is going out originating path 1 on port 1A2 gets looped so that it is output
to the WAN on through path 3 on shaping port 1A1.

Configuring PVCs
Switch B
Switch A
VP 3
1A1 shaping port WAN PVP Mesh

VP 2
switch fabric
Configuring PVCs
VP 0
VP 0
Switch C
1A2
VP 1 VP 1
looping port (orig. and term. paths VP 0 and VP 1)
Figure 1.15 - PVPs Coming in Port 1A1 from WAN and Looped through Port 1A2
Then, as shown in the example in Figure 1.15, the traffic that is coming in through path 2 on
port 1A1 from the WAN gets looped back to terminating path 0 on port 1A2. Similarly, the
traffic that is coming in through path 3 on port 1A1 from the WAN gets looped back to termi-
nating path 1 on port 1A2.
In your own network, you will repeat this process on these same two ports for as many paths
as you want to shape.
Before you configure the paths, you should

NOTE configure the network module that will contain
the looping port to maximize the number of
multicast connections. On a Series C network
module, use model number 5 under hardware
netmod traffic c modify -model. On a
Series LC network module, use model number 6,
7, or 8 under hardware netmod traffic lc
modify -model.

Configuring PVCs
To configure the shaped originating paths as shown in the example in Figure 1.14 and
Figure 1.15, perform the following steps:
1. First, create the UPC contract to be used on the through paths that are going to be
output on the shaping port on the Series D network module to the WAN.
myswitch:connections upc-> new -index 2 -servicecat cbr -pcr 42452 -gcra off
-schdmod smoothed
You must use the -gcra off option to disable

NOTE policing. This allow bursts without dropping
cells. The shaped rate is equal to PCR for CBR
traffic and equal to equivalent capacity for VBR
traffic. You must also use the smoothed
scheduling (-schdmod) option so that a rate
group is created when a PVC or PVP is
configured with this UPC.
2. Next, delete any existing path(s) on the looping port as follows:
myswitch:connections upc-> connections path term

myswitch:connections path term-> delete -atmif 1a2 -vpi 0
By default, path 0 is the only path that exists. If

NOTE path 1 already exists, you must delete it using
the command connections path term
delete -atmif 1a2 -vpi 1.
3. Recreate the paths on the looping port using the -loopvpi option to ensure that
cells from the WAN port get looped back into the terminating path.
myswitch:connections path term-> new -atmif 1a2 -vpi 0 -loopvpi 2

myswitch:connections path term-> new -atmif 1a2 -vpi 1 -loopvpi 3

Configuring PVCs
4. Create a PVP to the shaping port on a Series D network module and apply the UPC
contract that you created in step 1 so that it shapes the cells.
myswitch:connections path term-> ..

myswitch:connections path term-> through
myswitch:connections path through-> new -iatmif 1a2 -ivpi 2 -oatmif 1a1 -ovpi 2
-upc 2
-upc 2
Configuring PVCs
-upc 2
5. Create a PVP in the opposite direction and use -loopvpi to ensure that the cells
on each through path get looped back the terminating paths on the looping port
(1a2).
-loopvpi 0
-loopvpi 1
6. Now use a loopback command to loop the transmit side to the receive side on the
looping port (1a2).
myswitch:connections path through-> hardware port

myswitch:hardware port-> modify -hwportname 1a2 -adminstatus down
myswitch:hardware port-> sonet modify -hwportname 1a2 -loopback diag
Any required signalling paths must be recreated

NOTE at this point.
7. Additional shaped originating paths can be added later on these same two ports
as needed by repeating the steps in this section.

Configuring PVCs
1.5.2.2 Terminating a PVC at a Switch

Sometimes it is necessary to create a PVC between a host and a switch fabric that is at a remote
location. In this case, the PVC should be created from the host to the control port (CTL) of the
switch and vice versa.
Some additional configuration is necessary for communication to be established between the
host and the switch fabric. The switch needs an entry in its ATM ARP cache in order to send
cells destined for the host with the correct VPI and VCI and to pass received cells with a spe-
cific VPI and VCI to IP. This configuration can be done using AMI as shown in the following
subsections. (See the atmarp (8c) man page for more information.)
1.5.2.3 Creating ATM ARP Entries

To create a FORE IP PVC ARP entry, log in to AMI. Data on this PVC is encapsulated using
null encapsulation (also known as VC-based multiplexing) as specified in RFC-1483. Enter the
following parameters:
myswitch:services atmarp-> new-pvc

Usage:
[-ipaddress] <IP Address> IP Address
[[-conntype] <conn_type>] Connection Type (default: classicalIpPVC)
[[-vpi] <integer>] VPI
[[-vci] <integer>] VCI
[[-aaltype] (aal34|aal5)] AAL Type
[[-interface] <interface>] Interface (default: qaa0)
usage parameters.
Once the parameters are entered, the entry is created instantly by the SCP. At the same time, a
command is entered automatically into the current CDB which creates this ATM ARP entry
each time the SCP is restarted.

Configuring PVCs
1.5.2.4 Listing ATM ARP Entries

To verify that the ARP entries exist correctly for the outgoing PVC connection from the SCP to
the host, display the ATM ARP cache by logging in to AMI and entering the following param-
eters:
services atmarp show
IPaddress If VPI VCI AAL Type Direction

198.29.22.9 asx0 0 63 aal5 foreIpSVC outgoing
198.29.22.15 asx0 0 231 aal5 foreIpSVC pending
Configuring PVCs
198.29.22.37 asx0 0 65 aal34 foreIpSVC pending
IPaddress If NSAP Address
198.29.17.3 qaa0 0x47.0005.80.ffe100.0000.f21b.0138.002048102754.00
198.29.17.10 qaa0 0x47.0005.80.ffe100.0000.f21b.0137.002048100be6.00
198.29.17.15 qaa0 0x47.0005.80.ffe100.0000.f21b.0137.00204810048d.00
198.29.17.52 qaa0 0x47.0005.80.ffe100.0000.f21b.0138.0020481b0138.00
Field Description
IPaddress The IP address for this connection.
If The name of the IP interface for this connection.
VPI The virtual path number.
VCI The virtual channel number.
AAL The AAL type of the given connection.
Type Shows what kind of connection this is. Can be foreIpPVC, foreIpSVC,
classicalIpPVC, or classicalIpSVC.
Direction Outgoing means this is an outgoing connection.

Incoming means this is an incoming connection.
Pending means that a connection has not (yet) been established.
Incomplete means that the IP-to-ATM address mapping is not yet known for the given
IP address.
NSAP Address The NSAP address for this connection.

Configuring PVCs
1.5.3 Creating a Virtual Channel

To create a new virtual channel, log in to AMI and enter the following parameters:
myswitch:connections channel-> new

Usage:
[-iatmif] <AtmIf> Input AtmIf
[-ivpi] <integer> Input VPI
[-ivci] <integer> Input VCI
[-oatmif] <AtmIf> Output AtmIf
[-ovpi] <integer> Output VPI
[-ovci] <integer> Output VCI
[[-upc] <UPC Index>] UPC Contract (default: 0)
[[-ctype] <connection_type>] Connection Type
usage parameters.
The following is an example of how to create a uni-directional virtual channel which specifies
the call record connection type as transit:
myswitch:connections channel-> new -iatmif 3b1 -ivpi 0 -ivci 100 -oatmif 3b4
-ovpi 0 -ovci 100 -ctype tran
The following is an example of how to create a uni-directional virtual channel which has the
name “customer_a” assigned to it:
myswitch:connections channel-> new -iatmif 3b2 -ivpi 0 -ivci 145 -oatmif 3b3
-ovpi 0 -ovci 145 -name customer_a
The following is an example of how to create a bi-directional virtual channel on an ASX-1000

switch, ASX-1200 switch, or TNX-1100 switch. To create a VCC going in port 2A1, VPI 0, VCI
100 on the switch board installed in slot 2 and going out port 4B1, VPI 0, VCI 100 on the switch
board installed in slot 4, enter the following:
myswitch:connections channel-> new -iatmif 2a1 -ivpi 0 -ivci 100 -oatmif 2e4
-ovpi 0 -ovci 100

Configuring PVCs
myswitch:connections channel-> new -iatmif 2e4 -ivpi 0 -ivci 100 -oatmif 2a1
-ovpi 0 -ovci 100
myswitch:connections channel-> new -iatmif 4b1 -ivpi 0 -ivci 100 -oatmif 4e2
-ovpi 0 -ovci 100
myswitch:connections channel-> new -iatmif 4e2 -ivpi 0 -ivci 100 -oatmif 4b1
-ovpi 0 -ovci 100
In the first line in the first pair, notice that the output port is 2E4. This is the intra-fabric port.
The 2 means the connection is coming out of the switch board in slot 2 through the intra-fabric
Configuring PVCs
port. The E represents the intra-fabric port. The 4 means the connection is destined for switch
board in slot 4. 2E4 then becomes the input port in the second line.
In the first line in the second pair, notice that the output port is 4E2. This is the intra-fabric
port. The 4 means the connection is coming out of the switch board in slot 4 through the intra-
fabric port. The E represents the intra-fabric port. The 2 means the connection is destined for
switch board in slot 2. 4E2 then becomes the input port in the second line.
Once the parameters are entered, the virtual channel is created instantly by the SCP. At the
same time, a command is entered automatically into the current configuration database,
which means that this virtual channel is created every time the SCP is restarted.

Configuring PVCs
1.5.4 Creating SPVCs

A Smart Permanent Virtual Circuit Connection (SPVCC, hereafter referred to as SPVC) is a
connection that goes across multiple switch fabrics. An SPVC looks like a PVC at the local and
remote endpoints with an SVC in the middle. SPVCs are more robust than PVCs. If a link car-
rying a PVC goes down, then the PVC goes down. If a link carrying an SPVC goes down and
there is an alternate route, then the end switch fabrics of the SPVC automatically reroute the
SPVC around the failed link.
1.5.4.1 Creating a SPANS SPVC

To create a SPANS SPVC, you must configure both ends of the connection on the two switch
fabrics. This means you must have an AMI session open on both the source (local) switch fab-
ric and the destination switch fabric. You need to create a unidirectional SPANS SPVC in each
direction. Perform the following steps:
1. Log on to the two switches that will be the endpoints of your SPVC. In this exam-
ple, the switches are called shark and tuna.
2. Ensure that SPANS signalling is up on both switches by checking the Oper State
field under signalling spans show. If it is up, proceed to step 3. If it is down,
create a SPANS signalling channel on both switches on the ports you will be using
for the SPVC (in this example 1A1) using sig spans new.
3. To create a bidirectional SPANS SPVC, you need to create two unidirectional
SPVCs. Create the source side of the first unidirectional SPVC. For this unidirec-
tional SPVC, shark is the source and tuna is the destination as follows:
shark:connections smartchannel spans source-> new

Usage:
[-spvcid] <integer> Spvc Id
[-inatmif] <AtmIf> In AtmIf
[-invpi] <integer> In VPI
[-invci] <integer> In VCI
[-destspvcid] <integer> Dest Spvc Id
[-destswitchaddr] <SPANS Address> Dest Switch Addr
[[-allocbandwidth] <integer>] Alloc Bandwidth
See Part 1 of the AMI Configuration Commands

NOTE Reference Manual for descriptions of the usage
parameters.
shark:connections smartchannel spans source-> new -spvcid 1 -inatmif 1a1 -invpi

0 -invci 100 -destspvcid 1 -destswitchaddr 00000038.f21a3656

Configuring PVCs
On the destination switch (tuna), use the

NOTE command system show and copy and paste the
value from the SPANS Address field as the
-destswitchaddr value.
4. Now create the destination side of the first unidirectional SPVC. This time use the
SPANS Address from the source switch (shark) as the -destswitchaddr value
as follows:
tuna:connections smartchannel spans destination-> new
Configuring PVCs
Usage:
[-spvcid] <integer> Spvc Id
[-srcspvcid] <integer> Src Spvc Id
[-srcswitchaddr] <SPANS Address> Src Switch Addr
[-outatmif] <AtmIf> Out AtmIf
[-outvpi] <integer> Out VPI
[-outvci] <integer> Out VCI
[[-allocbandwidth] <integer>] Alloc Bandwidth

parameters.
tuna:connections smartchannel spans destination-> new -spvcid 1 -srcspvcid 1

-destswitchaddr 00000038.f21a333b -outatmif 1a1 -outvpi 0 -outvci 100
One unidirectional SPVC is now complete.

5. Verify that the unidirectional SPVC is up on both switches by looking at the State
field as follows:
shark:connections smartchannel spans source-> show

Local Remote State
ID AtmIf VPI VCI BW ID
1 1A1 0 100 0 1 up
Remote ATM Addr: 00000000.f21a3656
tuna:connections smartchannel spans destination-> show

Local Remote State
1 1A1 0 100 0 1 up
Source ATM Addr: 00000000.f21a333b

Configuring PVCs
6. Create another unidirectional SPVC going in the other direction. This time tuna is
the source and shark is the destination as follows:
tuna:connections smartchannel spans source-> new -spvcid 1 -inatmif 1a1 -invpi

0 -invci 100 -destspvcid 1 -destswitchaddr 00000038.f21a333b
shark:connections smartchannel spans destination-> new -spvcid 1 -srcspvcid 1

-destswitchaddr 00000038.f21a3656 -outatmif 1a1 -outvpi 0 -outvci 100
7. Verify that the second unidirectional SPVC is up on both switches by looking at the
State field as follows:
tuna:connections smartchannel spans source-> show

Local Remote State
1 1A1 0 100 0 1 up
Remote ATM Addr: 00000000.f21a333b
shark:connections smartchannel spans destination-> show

Local Remote State
1 1A1 0 100 0 1 up
Source ATM Addr: 00000000.f21a3656
Your bidirectional SPANS SPVC is now complete.

Configuring PVCs
1.5.4.2 Creating a PP PNNI SPVC

To create a point-to-point (PP) PNNI SPVC, perform the following steps:
2. Find the destination NSAP on the destination switch as follows:
tuna:connections-> destnsap -domainid 1 -atmif 1a1

Destination NSAP : 0x47.000580ffe1000000f21a3656.0020480d0000.00
Configuring PVCs
3. Create the PNNI PP SPVC on the source switch. Copy and paste the value from the
Destination NSAP field in the previous step as the -calledatmaddr value as
follows:
shark:connections smartchannel pnni pp-> new

Usage:
[[-index] <integer>] index (default: 1)
[-callingatmif] <AtmIf> Src AtmIf
[-callingvpi] <integer> Src VPI
[-callingvci] <integer> Src VCI
[-calledatmaddr] <NSAP Address> Destination NSAP
[[-calledvpi] <integer>] Called VPI
[[-calledvci] <integer>] Called VCI
[[-fwdupckey] <UPC Index>] Fwd UPC (default: 0)
[[-bckupckey] <UPC Index>] Bck UPC (default: 0)
[[-bearerclass] <bearerclass>] Bearer Class (default: classX)
[[-susceptclip] (no|yes)] CLIP (default: no)
[[-fwdqosclass] <QoS Class>] Fwd QoS (default: class0)
[[-bckqosclass] <QoS Class>] Bck QoS (default: class0)
[[-name] <text>] Name (default: “”)
[[-reroutestatus] (enabled|disabled)] Reroute Status (default: disabled)
[[-callingdomain] <integer>] Calling Domain (default: 1)
[[-qosindex] <integer>] QoS Index (default: 0)
[[-backoffstatus] (enabled|disabled)] Backoff Status (default: enabled)
[[-priority] <integer>] Priority (default: 5)
[[-dtltag] <integer>] DTL Tag (default: 0)
[[-autodtl] (enabled|disabled)] Auto DTL (default: enabled)
[[-rgroupid] <integer>] Redundancy Group ID
[[-secondaryvpi] <integer>] Secondary VPI
[[-secondaryvci] <integer>] Secondary VCI

Configuring PVCs

parameters.
shark:connections smartchannel pnni pp-> new -index 1 -callingatmif 1a1

-callingvpi 0 -callingvci 100 -calledatmaddr
0x47.000580ffe1000000f21a3656.0020480d0000.00 -calledvpi 0 -calledvci 100
4. Verify that the PNNI PP SPVC is up on the source switch by looking at the State
field as follows:
shark:connections smartchannel pnni pp-> show

INDEX Src: ATMIF VPI VCI UPC VPVC-SEL PRIORITY STATE
Dst: ATMIF VPI VCI UPC
-----------------------------------------------------------------------------
1 1A1 0 100 0 require 5 up
0 0 100 0
Destination: 0x47.0005.80.ffe100.0000.f21a.3656.0020480d0000.00
Your bi-directional PNNI PP SPVC from the source switch to the destination switch is now
complete.
When creating a PP SPVC, the -fwdqosclass

NOTE and -bckqosclass options must be either both
class0 or both a non-zero class. For example,
the -fwdqosclass cannot be class0 and the
-bckqosclass be class1. This combination is
invalid. Having an -fwdqosclass of class0
and a -bckqosclass of class0 is a valid
combination, and having an -fwdqosclass of
class1 and a -bckqosclass of class4 is a
valid combination, but having an
-fwdqosclass of class2 and a
-bckqosclass of class0 is not valid.

Configuring PVCs
1.5.4.3 Creating a PMP PNNI SPVC

The following is an example of how to create a simple PNNI PMP SPVC.
1. Log on to the various switches that will be the endpoints of your SPVC. In this
example, the root (source) switch is called shark and the party (destination)
switches are called tuna and flounder. You may have more parties than shown in
this example.
2. Create the root of the PMP SPVC on the source switch as follows:
shark:connections smartchannel pnni pmp root-> new ?
Configuring PVCs
[[-spvccid] <integer>] PMP SPVCCID (default: 2)
[-vci] <integer> VCI
[[-upc] <integer>] UPC (default: 0)
[[-bearerclass] <bearerclass>] Bearer Class (default: classX)
[[-susceptclip] (no|yes)] CLIP (default: no)
[[-fwdqosclass] <QoS Class>] Forward Qos (default: class0)
[[-priority] <integer>] Priority (default: 10)
[[-nextpartyindex] <integer>] Next Party Index
[[-domainid] <integer>] Domain ID (default: 1)
[[-rgroupid] <integer>] Redundancy Group ID
[[-secondaryvpi] <integer>] Secondary VPI
[[-secondaryvci] <integer>] Secondary VCI

parameters.
shark:connections smartchannel pnni pmp root-> new -spvccid 1 -atmif 1a1 -vpi 0 -vci 32

Configuring PVCs
3. Create the parties. First, find the destination NSAP on one of the destination party
switches as follows:
tuna:connections-> destnsap ?
Usage:
[-domainid] <unsigned> Domain ID

Destination NSAP: 0x47.000580ffe1000000f21c3950.0020480d00bc.00
4. Go back to the source switch and create the party. Copy the NSAP address from the
previous step into the command for the -atmaddr <NSAP Address> parameter
as follows:
shark:connections smartchannel pnni pmp root-> up
shark:connections smartchannel pnni pmp-> party

shark:connections smartchannel pnni pmp party-> new ?
[-spvccid] <integer> PMP SPVCCID
[[-partyid] <integer>] Party ID
[-atmaddr] <NSAP Address> Party NSAP Address
[[-vpi] <integer>] Called VPI
[[-vci] <integer>] Called VCI
[[-name] <text>] Party Name
[[-ftdtlindex] <integer>] FT-PNNI DTL
[[-reroutestatus] <reroutestatus>] Reroute Status (default:
rerouteNotRequested)
[[-qosindex] <integer>] QoS Exp Index (default: 0)
[[-autodtl] (enabled|disabled)] Auto DTL (default: enabled)
shark:connections smartchannel pnni pmp party-> new -spvccid 1 -partyid 1

-atmaddr 0x47.000580ffe1000000f21c3950.0020480d00bc.00

Configuring PVCs
5. Create the other party. Find the destination NSAP on the other destination party
switch as follows:
flounder:connections-> destnsap ?
Usage:
[-domainid] <unsigned> Domain ID
flounder:connections-> destnsap -domainid 1 -atmif 4c2

Destination NSAP: 0x47.0005.80.ffe100.0000.f21c.3950.0020480d0001.80
Configuring PVCs
6. Go back to the source switch and create the party. Copy the NSAP address from the
previous step into the command for the -atmaddr <NSAP Address> parameter
as follows:
shark:connections smartchannel pnni pmp party-> new -spvccid 1 -partyid 10

-atmaddr 0x47.0005.80.ffe100.0000.f21c.3950.0020480d0001.80
7. You can verify the information as follows:
shark:connections smartchannel pnni pmp party-> show

SPVCCID PartyId VPI VCI State VPVCSel QosExpIndex Time
1 1 0 33 up noPref 0 1 days 19:27
PartyNsapAddress: 0x47.0005.80.ffe100.0000.f21c.3950.0020480d00bc.00
1 10 0 34 up noPref 0 00:00
PartyNsapAddress: 0x47.0005.80.ffe100.0000.f21c.3950.0020480d0001.80
If you have more parties than shown in the example, you can repeat step 5 and step 6 for each
additional party. Your PMP PNNI SPVC is now complete.

Configuring PVCs
1.6 Traffic Types
The following sections contain informal

NOTE definitions of the concepts presented. For a
detailed understanding of these issues, please
see the ATM Forum’s UNI 3.1 and TM 4.0
Specifications.
Quality of Service (QOS) Management is based on the bandwidth parameters associated with
a virtual connection and the class of service and ATM Adaptation Layer (AAL) used for that
connection. In order to support voice, video, and data, the ATM Forum has defined classes of
service, or traffic types: Constant Bit Rate (CBR), real time Variable Bit Rate (rtVBR), non-real
time Variable Bit Rate (nrtVBR), Available Bit Rate (ABR), and Unspecified Bit Rate (UBR).
• At connection set-up time, traffic that uses a CBR parameter, such as a voice sig-
nal, makes a request for a dedicated Peak Cell Rate (PCR). Once the PCR is
defined, the ATM network must be able to guarantee that amount of bandwidth
for the duration of the connection.
• At connection set-up time, traffic that uses a rtVBR or nrtVBR parameter, such as a
video and data, makes a request for a dedicated PCR, Sustainable Cell Rate (SCR),
and Maximum Burst Size (MBS). Once these cell rates are defined, the ATM net-
work must be able to guarantee these rates for the duration of the connection.
• At connection set-up time, ABR traffic makes a request for a dedicated PCR and
Minimum Cell Rate (MCR). Once these cell rates are defined, the ATM network
must be able to guarantee the MCR rate for the duration of the connection. ABR
traffic sources adjust their transmission rate in response to information they
receive describing the status of the network and its capability to successfully
deliver data.
• UBR traffic, such as broadcast information and ARP messages, is also known as
“best effort” service. UBR provides no bandwidth guarantees.
Because ATM is designed to provide a single network to transport this variety of traffic
classes, FORE’s traffic policing and Connection Admission Control (CAC) schemes are vital to
allowing this mix of traffic to flow smoothly.

Configuring PVCs
1.7 Traffic Policing (Usage Parameter Control)

Traffic policing, also known as Usage Parameter Control (UPC), is a method of ensuring fair
allocation of network resources and of assessing the cells entering the switch for conformance
with pre-established traffic bandwidth contracts. Those cells that exceed the specified contract
are “tagged” or “dropped,” depending on what is defined in the contract. This ensures that
the connections with reserved bandwidth are not exceeding their reservations. FORE Systems’
switches use a combination of “leaky bucket,” or Generic Cell Rate Algorithm (GCRA) hard-
ware and user-configurable parameters in AMI to perform these policing functions.
Configuring PVCs
1.7.1 Leaky Bucket Algorithm
The first important concept to understand is the leaky bucket algorithm. Leaky buckets are a
mechanism by which cells entering the port are monitored for compliance with UPC traffic
contracts that have been negotiated at connection set-up time. Before the leaky buckets are
discussed, it is important to understand the parameters that are being measured by the buck-
ets, as shown in Table 1.1. These parameters are informally defined as follows:
• Peak Cell Rate (PCR) - the maximum number of cells per second
• Cell Delay Variation Tolerance (CDVT) - the tolerance for variation in the inter-
arrival time of these cells, or the amount of jitter that can be accepted by the net-
work
• Sustainable Cell Rate (SCR) - the average rate of cell transmission for this connec-
tion, taking bursting into account
• Burst Tolerance (BT) - the maximum amount of cells that can be transmitted above
SCR
• Minimum Cell Rate (MCR) - the minimum rate that the network has to guarantee
for an ABR connection
The leaky bucket algorithm is basically a timer which assesses if cells entering the port con-
form to the parameters listed above. As a cell arrives, the timer assesses if the cell is on time,
late, or early. If the cell is determined to be on time or late (based on the traffic parameters), the
cell is allowed to pass unchanged. If the cell is early (which, in turn, causes the cell stream to
exceed the specified parameters), the cell is considered non-conforming and is either dropped
or tagged (the CLP bit is set to 1), depending on the specified contract.
The first bucket in this analogy measures the PCR, or the rate at which the bucket drains. It
also considers the CDVT, or the depth of the bucket. The second bucket measures the SCR, or
the rate at which the bucket drains, and the BT, or the depth of the second bucket.

Configuring PVCs
1.7.2 Non-conforming Cells: Tagging vs. Dropping

Second, it is important to understand the concept of tagging and dropping. Each ATM cell has
a Cell Loss Priority (CLP) bit which indicates if the network can drop it under congested con-
ditions. When the CLP bit is set to 0 (or CLP=0), the cell is assessed for compliance with traffic
parameters associated with the CLP=0 stream. If the traffic parameters dictate that non-com-
pliant cells should be “tagged,” the CLP bit is set to 1 (or CLP=1) by the UPC contract, which
means that upon experiencing congestion further in the network, these CLP=1 cells are
dropped in preference to CLP=0 cells.
Table 1.1 shows the various traffic contracts and user-configurable actions to be executed upon
the detection of non-conforming cells.
Table 1.1 - Summary of Traffic Contract Variables and Policing Actions
Policin Action on
Policing
g non-conforming cells
PCR SCR MBS Bucket 1 Bucket 2 on Switch
Schem
Fabric
e Bucket 1 Bucket 2
CBR PCR01 PCR01, Yes Drop

CDVT CLP=0+1
CBR0 PCR0, PCR01, PCR0, CDVT Yes Drop Drop CLP=0

PCR01 CDVT CLP=0+1
CBR0T PCR0, PCR01, PCR0, CDVT Yes Drop Tag CLP=0 (change bit
AG PCR01 CDVT CLP=0+1 to CLP=1)
VBR1 PCR01 SCR01 MBS01 PCR01, SCR01, Yes Drop Drop CLP=0+1
CDVT BT01+CDVT CLP=0+1
VBR2 PCR01 SCR0 MBS0 PCR01, SCR0, Yes Drop Drop CLP=0
CDVT BT0+CDVT CLP=0+1
VBR3 PCR01 SCR0 MBS0 PCR01, SCR0, Yes Drop Tag CLP=0 (change bit
CDVT BT0+CDVT CLP=0+1 to CLP=1)

Configuring PVCs
1.7.3 UPC Traffic Contract Parameters

The ATM Forum has defined different types of traffic contracts to be used in conjunction with
these leaky buckets. The parameters that make up these types of contracts are defined as fol-
lows:
• servcat - The service category of the contract (CBR, rtVBR, nrtVBR, ABR, UBR)
• pscheme - The policing scheme of the contract (CBR1, CBR0, CBR0TAG, VBR1,
VBR2, VBR3, ABR1, UBR1, UBR2)
• pcr - Peak Cell Rate (CLP01 for all police schemes)
Configuring PVCs
• scr - Sustained Cell Rate (CLP01 for VBR1, CLP0 for CBR0, CBR0TAG, VBR2, and
VBR3)
• mbs - Maximum Burst Size (CLP01 for VBR1, CLP0 for VBR2 and VBR3)
The specific combinations of these parameters that make up the ATM Forum contracts are
defined as follows:
new-cbr cbr1 <pcr>
new-cbr cbr0 <pcr> <scr>
new-cbr cbr0tag <pcr> <scr>
new-vbr (rtVBR|nrtVBR) (vbr1|vbr2|vbr3) <pcr> <scr> <mbs>
new-abr abr1 <pcr> <mcr>
The new-cbr cbr1 <pcr> contract is for CBR traffic. It only uses the first leaky bucket to assess
the conformance to PCR of the aggregate of the CLP=0 cells and the CLP=1 cells. Cells which
fail the PCR CLP=0+1 test are discarded.
The new-cbr cbr0 <pcr> <scr> and new-cbr cbr0tag <pcr> <scr> contract is for CBR traffic. It
uses the first leaky bucket to assess the conformance to SCR of the CLP=0 cells. It uses the sec-
ond leaky bucket to assess the conformance to PCR of the aggregate of the CLP=0 and the
CLP=1 cells. If the CBR0TAG policing scheme is set, the cells which fail the SCR CLP=0 test
are tagged as CLP=1 and passed on to the second leaky bucket to be tested for PCR CLP=0+1
conformance. Cells which fail the PCR test on CLP=0+1 are discarded. If the CBR0 policing
scheme option is set, cells which fail the SCR CLP=0 test are discarded and cells which fail the
PCR test on CLP=0+1 are discarded.
The new-vbr (rtVBR|nrtVBR) vbr1 <pcr> <scr> <mbs> contract is for VBR traffic.
The first leaky bucket assesses the conformance to PCR of the aggregate of CLP=0 cells and
the CLP=1 cells and the second leaky bucket assesses the conformance to SCR and BT of this
same combination. Cells which fail the PCR test are dropped. Cells which pass the PCR test,
but which fail SCR and BT test are dropped.

Configuring PVCs
The new-vbr (rtVBR|nrtVBR) (vbr2|vbr3) <pcr> <scr> <mbs> contract is for VBR traffic. It
uses the first leaky bucket to assess the conformance to PCR of the aggregate of the CLP=0 and
the CLP=1 cells. Cells that fail this test are discarded. It uses the second leaky bucket to assess
the conformance to SCR and BT of the CLP=0 cells. If the VBR3 policing scheme is set, the cells
which fail the SCR and BT CLP=0 test are tagged as non-conforming cells. If the VBR2 policing
scheme is set, cells which fail the SCR and BT CLP=0 test are discarded.
The new-abr abr1 <pcr> <mcr> contract is for ABR traffic. It only uses the first leaky bucket to
assess the conformance to PCR of the aggregate of the CLP=0 cells and the CLP=1 cells. Cells
which fail the PCR CLP=0+1 test are discarded. The MCR is the guaranteed minimum rate
and the PCR is the maximum required rate. The MCR may be set to 0, which indicates “best
effort” service. If MCR is set greater than 0, then the rate is guaranteed, up to MCR. However,
between MCR and PCR, cells may be dropped when congestion is experienced.
1.7.4 AMI UPC Commands

AMI allows you to create a UPC contract using these combinations of traffic parameters. You
can use the following syntax to create a UPC contract in AMI or you can use the new-cbr,
new-abr, new-ubr, or new-vbr commands. See Chapter 3 in Part 1 of the AMI Configuration
Commands Reference Manual for the syntax for those commands.
myswitch:connections upc-> new
Usage:
[-index] <integer> Index
[[-servcat] <ServCat>] Service Category (default: UBR)
[[-pscheme] <pScheme>] Policing Scheme (default: ubr1)
[[-pcr] <integer>] PCR (default: “”)
[[-scr] <integer>] SCR (default: “”)
[[-mbs] <integer>] MBS (default: “”)
[[-cdvt] <integer>] CDVT
[[-aal5epd] (yes|no)] AAL5 EPD (default: no)
[[-gcra] (on|off)] GCRA Policing (default: on)
[[-aal5] (yes|no)] AAL5 (default: no)
[[-pktdisc] (on|off)] Packet Discard (default: on)
[[-pppol] (on|off)] PP Policing (default: off)
[[-schdmod] <SchedMode>] Scheduling Mode (default: default)
[[-altclp] (on|off)] Alt CLP Threshold (default: off)
[[-mcr] <integer>] ABR MCR (default: “”)
[[-aal5cnt] <CountMode>] AAL5 Counting Mode (default: default)
[[-subcat] <integer (1..16)>] Service Sub-category (default: default)
[[-lbmincr] <integer>] Load Balanced MinCR
[[-lbpcr] <integer>] Load Balanced PCR

Configuring PVCs

parameters.
Remember, when you create a UPC contract, it is

NOTE not actually used until you assign it to a VPC,
VCC, or SPANS path. UPC contracts are not
assigned to VPTs.
Configuring PVCs
The following is an example of how to create a UPC contract:
myswitch:connections upc-> new -index 5 -servcat vbr -pcr 500 -scr 200 -mbs 250 -cdvt
1000 -aal5 yes -pppol on -name vbr_upc
This example specifies a contract named “vbr_upc”, which is a VBR contract with an index of
5, a PCR of 500 cells/sec (or kbps), an SCR of 200 cells/sec (or kbps), an MBS of 250 cells (or
kilobits), a CDVT of 1,000 microseconds, and partial packet policing enabled.
For more information regarding traffic contracts,

NOTE please refer to the ATM Forum TM 4.0
Specification.
PVCs that use UPC contracts that contain any of

NOTE the -gcra, -aal5, -pktdisc, and -pppol
options are valid only when the following
commands are set to svcOn or svcOff:
• interfaces atmif modify -gcracbr
-gcranrtvbr -gcrartvbr
• interfaces atmif modify
-aal5pktdcbr -aal5pktdnrtvbr
-aal5pktdrtvbr -aal5pktdubr
• interfaces atmif modify -pppcbr
-pppnrtvbr -ppprtvbr
• interfaces atmif modify -ubrtagging
Use interfaces atmif show -tm and
interfaces atmif show -advancedtm to
check these settings.

Configuring PVCs

2.1 Introduction
This chapter describes how to design, configure, and maintain a Classical IP ATM network.
The term classical indicates that the ATM network has the same properties as existing legacy
LANs. That is, even though ATM technology allows for large, globally connected networks,
for example, it is only used in the LAN environment as a direct replacement of existing LAN
technology. The classical model of LANs connected through IP routers is maintained in ATM
networks. RFC-1577 provides the standard for Classical IP over ATM.
Classical IP over ATM is different than IP in legacy LANs in that ATM provides a virtual con-
nection environment through the use of Permanent Virtual Circuits (PVCs) and/or Switched
Virtual Circuits (SVCs). SVC management is performed via the ATM Forum UNI 3.0 Specifica-
tion, which specifies Q.2931. Q.2931 is a broadband signalling protocol designed to establish
Configuring Classical
connections dynamically at the User-Network Interface (UNI). Q.2931 uses Service Specific
Connection Oriented Protocol (SSCOP) as a reliable transport protocol, and all signalling
occurs over VPI: 0, VCI: 5. Q.2931 connections are bidirectional, with the same VPI/VCI pair
used to transmit and receive.
IP
Once a Classical IP connection has been established, IP datagrams are encapsulated using
IEEE 802.2 LLC/SNAP and are segmented into ATM cells using ATM Adaptation Layer type
5 (AAL5). Additionally, the default Maximum Transmission Unit (MTU) is 9,180 bytes (the
SNAP header adds 8 more bytes) with a maximum packet size of 65,535 bytes. There is cur-
rently no support for IP broadcast datagrams or IP multicast datagrams in a Classical IP envi-
ronment.

Configuring Classical IP
2.1.1 Logical IP Subnets

An important concept in Classical IP networks is that of a Logical IP Subnet (LIS). An LIS is a
group of hosts configured as members of the same IP subnet (that is, they have the same IP
network and subnetwork numbers). In this sense, one LIS can be equated to one legacy LAN.
It is possible to maintain several overlaid LISs on the same physical ATM network. Therefore,
in a Classical IP ATM network, placing a host on a specific subnet is a logical choice rather
than a physical one. In this type of environment, communication between hosts in different
LISs is only permitted by communicating through an IP router which is a member of both LISs
(as per RFC-1577).
The number of LISs, and the division of hosts into each LIS, is purely an administrative issue.
Limitations of IP addressing, IP packet filtering, and administrative boundaries may guide a
manager into establishing several LISs onto a single ATM network. Keep in mind, though,
that communication between LISs must occur through IP routers.
2.1.2 Classical IP Interfaces

In order to support routing between multiple LISs, the switch software allows a switch to be
configured as a member of (and a router between) up to four distinct LISs. (The host adapter
software allows a host to be configured as a member of (and a router between) up to 16 dis-
tinct LISs.) Each LIS membership is through a separate Classical IP network interface. Existing
system level IP routing configuration tools are used to control routing through each of the
Classical IP interfaces in the same manner as routing among several physical interfaces. Even
though each Classical IP interface associated with a given physical interface uses the same
physical hardware, they are each configured separately with their own MTU, IP address, and
ATM address.
By default, the name of each of the Classical IP interfaces on a switch begins with qa. (On a
host, each of the Classical IP interfaces begins with ci and is user-configurable.) On a switch,
all of the Classical IP interfaces associated with physical unit zero have a as the next letter. All
of the Classical IP interfaces associated with physical unit one have b as the next letter, and so
forth. Finally, each Classical IP interface has its interface number as a suffix. As an example of
the above naming convention for switches, the name of the third Classical IP interface (unit 2)
on physical unit one is qab2.

2.1.3 SPANS Interface

While each of the Classical IP interfaces for a given physical interface is designed to support
Classical IP using Q.2931 signalling, a SPANS interface also exists for each physical interface.
The SPANS interface is asx0 in switch software and is fa0 in host adapter software.
The SPANS interface supports FORE IP on top of SPANS signalling. FORE IP allows commu-
nication using AAL4 or AAL5 with no encapsulation, uses a broadcast ARP for SPANS
address resolution, and supports direct communication of all hosts on a physical ATM net-
work without the use of IP routers. Since SPANS and Q.2931 signalling use different VCIs, a
host can simultaneously support FORE IP over SPANS as well as Classical IP over Q.2931 on
the same physical interface.
As a result of standard IP routing, all traffic sent out a SPANS interface uses FORE IP, while all
traffic sent out a Classical IP interface uses Classical IP. The Classical IP interfaces are qaaX
(where X is 0 - 3) in switch software and are ci in host adapter software.
Each of the SPANS interfaces should be assigned an IP address on a subnet different than the
subnets of any of the Classical IP interfaces. It is permissible to place multiple SPANS inter-
faces on the same subnet, and the driver load balances connections across these interfaces.
It is only necessary to configure the SPANS and Classical IP interfaces if the specific service
provided by that interface is required. A host sending only Classical IP would not need to con-
figure the SPANS interfaces. Likewise, a host sending only FORE IP would not need to config-
ure the Classical IP interfaces. Both the SPANS and Classical IP interfaces may be configured
IP
simultaneously, but they must be in separate subnets. Remember that Classical IP specific con-
figuration changes can only be done with the Classical IP devices, while SPANS specific con-
figuration changes can only be done with the SPANS devices.

2.2 Address Registration and ILMI

Before a host can establish connections over a physical interface, the host must know the
NSAP address for that interface. The primary purpose of Interim Local Management Interface
(ILMI) is to discover and register these NSAP addresses dynamically.
2.2.1 NSAP Addresses

For private ATM networks, addresses uniquely identify ATM endpoints. The UNI 3.0 address
format is modeled after that of an OSI Network Service Access Point, hence the name NSAP
address.
Three address formats have been specified: DCC, ICD, and E.164. FORE implements the ICD
ATM format. Per the UNI 3.0 specification, all private networks should accept initial call set-
up messages containing ATM addresses with any of the approved formats and forward the
calls as necessary.
An NSAP address consists of the following:
• a 13-byte network-side prefix - The prefix is the NSAP prefix of the switch to
which the host is attached.
• a seven-byte user-side part - This consists of the following:
- a six-byte End System Identifier (ESI) - The ESI is the unique IEEE MAC
address of the interface.
- a one-byte selector - Although each Classical IP interface for a given physical
interface uses the same prefix and ESI, the selector field is the part that indi-
cates the number of the specific Classical IP interface. On a switch, the selec-
tor field is 00 for qaa0, 01 for qaa1, 02 for qaa2, and 03 for qaa3.

2.2.2 Operating with ILMI Support

FORE Systems switches running ForeThought software provide support for ILMI. If ILMI is
supported on all of the switches and hosts in a given network, when a switch boots up, ILMI
enables the switch to discover all of the hosts attached to it and to send its ATM prefix associ-
ated with the port to those hosts dynamically. In return, the host prepends that prefix to its ESI
and selector fields, forming a complete ATM address. The host then notifies the switch of its
complete ATM address. These registration messages are sent and received over AAL5 using
VPI: 0, VCI: 16. Once ILMI registration has been completed, then connection setup can occur.
If a host changes network ports after an ATM address has been registered for its interface, all
existing connections are closed. If the new port is on a different switch, a new ATM address
(with a different network address prefix) is registered. The host can then begin to establish
new connections.
2.2.3 Operating without ILMI Support

If ILMI is not supported on a particular switch or host in a given network, then the ATM
addresses must be manually configured. If another vendor’s switch does not support ILMI, it
can not supply an ATM prefix to the hosts. Therefore, the user must assign a unique, valid pre-
fix to the switch. Additionally, the same prefix should be used for all hosts attached to that
switch.
IP
On the host, uniconfig is used to configure the ATM address for a specific interface. The
switch directly attached to this interface is then informed of this ATM address/port combina-
tion through commands in the ATM Management Interface (AMI). Once the host and network
have both been informed of this ATM address/port pair, the host may begin signalling.
2.2.4 Configuration
The choice to use ILMI for address registration is made at software installation time. Since
ILMI uses SNMP as its management protocol, the use of ILMI is tied into snmpd. The choice
can be made to run FORE’s SNMP agent and use ILMI (snmpd), run FORE’s SNMP agent
without using ILMI (snmpd -n), or just use ILMI (snmpd -i or ilmid -i).

2.3 ARP and ARP Servers
2.3.1 Theory
In order for a host to establish a connection to another host, it must first determine the other
host’s ATM address. ATM ARP (ATM address resolution protocol) is the procedure used to
resolve an IP address into an ATM address. Since the ATM standards do not currently support
broadcast on an ATM LAN, address resolution is performed by direct communication with a
special ARP server, rather than broadcasting ARP requests as is done in legacy LANs. Each
LIS must have only one ARP server configured, but a single ARP server can be the server for
several LISs.
Each host in an LIS must be configured with the ATM address of the host providing ARP ser-
vice for its LIS. On a switch, the ATM address of the ARP server can be obtained by using the
AMI command services atmarp arpserver show. On a host, the ATM address of the ARP
server can be obtained by using clipconfig show (remember to use the interface associated
with the given LIS).
On a switch, the ATM address of the ARP server can be configured by using the instructions
found in Section 2.3.2 of this manual. On a host, the ARP server address is configured into the
host at installation time using the configure_atm script.
Since only one ARP server can be functioning at a time in a given LIS, and since the ARP
server’s address is configured into each host, it is not possible to use multiple, redundant ARP
servers to improve robustness. If an ARP server becomes nonfunctional, a new ARP server
must be configured, and then each host within the LIS must be configured to use the new ARP
server. To configure a new ARP server address on a switch, use the the instructions found in
Section 2.3.2 of this manual. To configure a new ARP server address on a host, you must delete
the interface and recreate it.

2.3.2 Configuring a FORE Switch to be an ARP Server

FORE’s ATM switches also have the capability of being an ARP server. (FORE’s ATM adapters
can not be configured as an ARP server.) To configure a FORE ATM switch as an ARP server,
perform the following steps on only one of the SCPs:
1. On one of the SCPs, determine the ATM address of that SCP for the relevant inter-
face (qaa0 -> qaa3) using the following AMI command:
services atmarp getnsap -ifname
For example:
services atmarp getnsap -ifname qaa0
qaa0 NSAP address: 47000580ffe1000000f12400de0020481900de00
2. Set the NSAP address of the ARP server to be the ATM address of the interface that
you displayed in step 1 using the following AMI command:
services atmarp arpserver modify [-ifname] [-serverid] [-addr]
For example:
IP
services atmarp arpserver modify -ifname qaa0 -serverid 1 -addr
0x47000580ffe1000000f12400de0020481900de00
3. Set the ATM address of the ARP server on each of the other switches that will use
that switch as the ARP server using the same command found in step 2.

2.3.3 Classical IP Operation

Once a host knows its own ATM address and the ATM address of its ARP server it attempts to
establish a connection to the ARP server, which is used to send ARP requests and receive ARP
replies. When the connection to the ARP server has been established, the ARP server sends an
inverse ARP (InARP) request on the new VC to learn the host’s IP address. When an InARP
reply is received, the ARP server places that host’s IP address to ATM address mapping in its
ARP cache. Therefore, over time, the ARP server dynamically learns the IP-to-ATM address
mappings of all the hosts in its LIS. It can then respond to ARP requests directed toward it for
hosts in its LIS.
In order for a host to communicate with an ARP

NOTE server, it must have learned its own ATM
address and have been configured with the ATM
address of the ARP server.
A host can not resolve the ATM addresses of

other hosts in its LIS unless it can communicate
with its ARP server.
Since there is no mechanism for ARP servers to

exchange mapping information with each other,
it is imperative that each LIS be configured with
only one ARP server.
When a host wants to communicate with another host in its LIS, it first sends an ARP request
to the ARP server containing the IP address to be resolved. When an ARP reply is received
from the ARP server, the host creates an entry in its ARP cache for the given IP address and
stores the IP-to-ATM address mapping. This ARP cache entry is marked as complete. To
ensure that all of the IP-to-ATM address mappings known by a certain host are up-to-date,
hosts are required to age their ARP entries. A host must validate its ARP entries every 15 min-
utes (20 minutes on an ARP server). Any ARP entries not associated with open connections
are immediately removed.
A host validates its SVCs by sending an ARP request to the ARP server. A host validates its
PVCs, and an ARP server validates its SVCs, by sending an InARP request on the VC. If a
reply is not received, the ARP entry is marked invalid. Once an ARP entry is marked invalid,
an attempt is made to revalidate it before transmitting. Transmission proceeds only when val-
idation is successful. If a VC associated with an invalid ARP entry is closed, the entry is
removed.

2.3.4 Operational Issues

Certain hosts in an LIS may not support Classical IP. It is still possible to communicate with
these hosts (and for these hosts to communicate with one another) by using static ARP entries.
If a host does not support Classical IP, its IP-to-ATM address mapping should be placed in its
ARP server’s cache as a static entry. This allows other hosts that do support Classical IP to
contact their ARP server as usual and obtain the correct address mapping. If a host that does
not support Classical IP wants to initiate connections, the IP-to-ATM address mappings of the
destination hosts should be put in its ARP cache, again as static entries. By using static ARP
entries in the above fashion, the ability for all hosts to communicate is maintained.
There are some restrictions on the number of hosts that can be maintained dynamically. They
are as follows:
• In the default configuration, a host can only have approximately 250 virtual con-
nections open simultaneously. This means that an ARP server can only serve 250
clients, since each client must maintain a connection with its ARP server. This
may be a limitation if the ARP server is servicing multiple LISs.
• It is possible to increase the number of connections using AMI.
• Some hosts may be limited to supporting a maximum of 1,024 connections per
adapter.
IP

2.4 Classical IP PVCs
2.4.1 Theory and Configuration

Normally, ATM connections in a Classical IP environment are established dynamically using
UNI 3.0 or UNI 3.1. ARP, ILMI, and UNI 3.0 or UNI 3.1 all work together as described previ-
ously to set up an SVC. If a host from another vendor does not support Classical ARP or ILMI,
it is still possible to set up an SVC using work-arounds. If a host or a switch in an LIS does not
support UNI 3.0 or UNI 3.1, however, it is not possible to establish an SVC. In this case, a Clas-
sical IP PVC can be used for communication.
On each of the hosts, cliparp add -pvc is used to establish the PVC. An unused VPI/VCI
pair must be chosen for each host. PVCs using the chosen VPI/VCI pairs must also be set up
from each of the hosts to their connecting switch, and then on all of the switches between the
two connecting switches.
Both the incoming and outgoing connections are

NOTE set up simultaneously on the host, but they must
be set up individually on the switches. The same
VPI/VCI pair is used by a host to send on the
PVC as well as receive on the PVC. The IP
datagrams are sent over the PVC using AAL5
with LLC/SNAP encapsulation.
2.4.2 Revalidation and Removal

Normally, the device driver periodically checks that its PVCs are still established and func-
tioning. A host revalidates a PVC by sending InARP requests over the PVC, if the user speci-
fies that revalidation should occur by using the -reval option to cliparp add -pvc at the
time the PVC is created (e.g., -reval 10 means revalidation will occur every 10 minutes). If
the equipment attached to the FORE equipment supports revalidation, the user must choose
the -reval option. If an InARP reply is not received, the revalidation fails, the PVC is marked
invalid (as shown through cliparp show), and communication over the PVC is no longer
possible.
Once a PVC is marked invalid, an attempt is made to validate the PVC before transmitting.
Transmission proceeds only when validation is successful. It is possible to disable this revali-
dation feature by not specifying the -reval option. This is often desirable when the remote
end of the PVC (such as a video camera) does not support InARP.
A Classical IP PVC is removed on the host side using cliparp delete -pvc. Both the
incoming and outgoing connections are removed simultaneously. The PVC must then be
removed from each of the network switches involved.

2.5 Configuring the Network

In an ATM network, before any connections can be made, the two parties must know each
other’s ATM address in order to set up that connection.
To allow those connections to work, the ideal scenario is for all hosts and switches in the net-
work to have support for both ILMI and for RFC-1577 (Classical IP over ATM). However,
when using other vendors’ equipment, this may not be the case. This section describes how to
configure a network with the following scenarios:
• Configuring a third-party host that has no ILMI and no RFC-1577 support
• Configuring a third-party switch that has ILMI support, but no RFC-1577 support
• Configuring a third-party switch that has no ILMI support, but has RFC-1577
support
IP

2.5.1 Third-Party Host with No ILMI and No RFC-1577 Support

To configure a network with a third-party vendor’s host (or an edge device) that supports nei-
ther ILMI nor RFC-1577 (as shown in Figure 2.1), perform the following steps:
FORE Switch
FORE
FORE Switch
(ARP server)
FORE Third-Party Host

(no ILMI, no RFC-1577)
Figure 2.1 - Configuring a Third-Party Host with No ILMI and No RFC-1577 Support
1. Before beginning this process, be sure that ForeThought software is installed and
running on the FORE equipment.
2. Using the configuration software of the third-party host, assign that host an NSAP
address that has the same prefix as the switch fabric to which it is connected.
3. Configure the switch that is the ARP server so that it has a static route to the third-
party host using the following AMI command:
routing ftpnni staticroute new
Be sure to use a host mask value of 152.

2.5.2 Third-Party Switch with ILMI and No RFC-1577 Support

To configure a network with a third-party vendor’s switch that supports ILMI, but not
RFC-1577, (as shown in Figure 2.2), perform the following steps:
FORE FORE
Switch A Switch B
Third-Party Switch
ILMI, no RFC-1577
IP
= FORE Systems host
Figure 2.2 - Configuring a Third-Party Switch with ILMI Support and No RFC-1577
1. Be sure that ForeThought software has been installed on all of the hosts and that
ILMI was set in the process. ILMI dynamically performs address registration for
all of the hosts.
2. Configure a static ATM route to the third-party switch on FORE switch “B” that is
physically connected to the third-party switch using the following AMI command:
Be sure to use a network mask value of 104.

3. Configure two static NSAP routes on the third-party switch, one to each of the
FORE switches to which the third-party switch is connected, using the third-party
vendor’s configuration software. Be sure to use a network mask value of 104.

2.5.3 Third-Party Switch with RFC-1577 and No ILMI Support

To configure a network with a third-party vendor’s switch that does not support ILMI, but
does support RFC-1577 (as shown in Figure 2.3), perform the following steps:
FORE FORE
Switch A Switch B *
Third-Party Switch
RFC-1577, no ILMI
*
* *
= FORE Systems host
Figure 2.3 - Configuring a Third-Party Switch with RFC-1577 and No ILMI Support
1. Be sure that ForeThought software has been installed on all of the FORE hosts and
that ILMI was set in the process. ILMI dynamically performs address registration
for all of the FORE hosts and FORE switches.
2. Statically configure the non-FORE (*) hosts with ATM addresses (edit the firmware
download script), using the same switch prefix for all of the hosts.
3. Configure a static ATM route to the third-party switch on FORE switch “B” that is
physically connected to the third-party switch using the AMI command:
Be sure to use a network mask value of 104. Also, be sure to use the same prefix
that was used to configure the hosts.
4. Configure two static ATM routes on the third-party switch, one to each of the
FORE switches to which the third-party switch is connected, using the third-party
vendor’s configuration software. Be sure to use a network mask value of 104.

3.1 Introduction
This chapter describes how to design, configure, and maintain an Emulated LAN (ELAN)
over an ATM network. An ELAN provides communication of user data frames among all
members of the ELAN, similar to a physical LAN. One or more ELANs may run simulta-
neously (and independently) on the same ATM network.
Each ELAN is composed of a set of LAN Emulation Clients (LECs), a LAN Emulation Config-
uration Server (LECS), and at least one LAN Emulation Server (LES) and Broadcast and
Unknown Server (BUS) pair (also referred to as colocated BUS or an intelligent BUS). In the
current software release, the LECS may reside either in a ForeRunner ASX-200WG,
ASX-200BX, ASX-1000, ASX-1200, ASX-4000, LE 25, LE 155, ESX-3000, TNX-210, or TNX-1100
switch, or in a UNIX workstation running Solaris 2.5, 2.5.1, or 2.6. The LES/BUS pair may
reside either in a PowerHub 7000, PowerHub 8000, ASN-9000, ASX-200WG, ASX-200BX,
ASX-1000, ASX-1200, ASX-4000, LE 25, LE 155, ESX-3000, TNX-210, or TNX-1100 switch, or in
a UNIX workstation running Solaris 2.5, 2.5.1, or 2.6. An additional software feature is Distrib-
uted LAN Emulation (DLE), which provides load sharing and fault tolerance to the ELAN.
The current software release supports the emulation of both Ethernet (IEEE 802.3) and Token
Ring ELANs.
3.1.1 Ethernet ELANs
Configuring an
The current software release supports emulation of Ethernet (IEEE 802.3) ELANs. In the cur-
Emulated LAN
rent release, each LEC resides on an ATM host system (PC, Macintosh, UNIX workstation,
ATM switch, PowerHub 7000, or ES-3810).
3.1.2 Token Ring ELANs

The current software release supports emulation of Token Ring (IEEE 802.5) ELAN services. In
the current release, Token Ring LECs can not be created on the switch. Token Ring LECs can
only be created on some hosts.

Configuring an Emulated LAN
3.2 ELAN Components

The components of an ELAN include LECs, and LAN Emulation services consisting of a
LECS, a LES, and a BUS. Although the ATM Forum specification allows the LES and BUS to be
located on different devices, more intelligent traffic handling is possible when they are located
on the same device. ForeThought 5.0.x or greater software requires the LES and BUS be
co-located (reside on the same device).
The LECS may reside in the same physical system as the LES/BUS or in a separate physical
system. For example, the LECS could reside in a switch, while the LES/BUS reside in a work-
station. Currently, the LECS is supported only on ForeRunner and TNX switches and on sys-
tems running Solaris. The LES/BUS are supported only on ForeRunner and TNX switches, on
PowerHub 7000s, on PowerHub 8000s, on ASN-9000s, and on systems running Solaris. The
functional interconnections of a simple ELAN consisting of two LECs, an LECS, a LES, and a
BUS are shown in Figure 3.1.
Workstation LAN Emulation Switch

Configuration Server or
(LECS) Bridge
LAN Emulation Server

(LES)
LAN Emulation LAN Emulation

Client Client
(LEC) Broadcast and Unknown (LEC)
Server
(BUS)
LAN Emulation Services Legacy

LAN
Figure 3.1 - Basic Emulated LAN Interconnections

3.2.1 LAN Emulation Client (LEC)

The LEC is the component in an end system that performs data forwarding, address resolu-
tion, and other control functions when communicating with other components within the
ELAN. It also provides a MAC level emulated Ethernet or Token Ring interface and appears to
higher level software as though a physical interface is present. Each LEC must register with
both the LES and BUS associated with the ELAN it wishes to join before it may participate in
the ELAN. To participate in multiple ELANs, an end system must have multiple LECs. Cur-
rently, up to 16 LECs are supported on ES-3810 switches and on adapter cards running Solaris
or Windows/NT software, and up to 4 LECs on adapter cards running Windows/95.
3.2.2 LAN Emulation Configuration Server (LECS)

The LECS is responsible for the initial configuration of LECs. It provides information about
available ELANs that a LEC may join, together with the address of the LES associated with
each ELAN. Using DLE in ForeThought 5.0.x and greater, the user may also configure the LECS
to associate multiple LES/BUS pairs with a given ELAN. This feature allows LECs to use a
single, anycast address to reach one of the other DLE peer servers for their ELAN if their local
server goes down. Normal address resolution through ForeThought PNNI, ATM Forum PNNI,
or IISP will locate the closest, active LES which is using the anycast address.
3.2.3 LAN Emulation Server (LES)

The LES implements the control coordination function for the ELAN. The LES provides the
service of registering and resolving MAC addresses to ATM addresses. A LEC registers its
own address with the LES. A LEC also queries the LES when the client wishes to resolve a
MAC address to an ATM address. The LES either responds directly to the client or forwards
the query to other clients so they may respond. There may be more than one instance of an
active LES per ELAN.
Configuring an
Emulated LAN
3.2.4 Broadcast and Unknown Server (BUS)
Unlike traditional shared-media LAN architectures such as Ethernet or Token Ring, ATM is
connection based. Therefore, it has no built-in mechanism for handling connectionless traffic
such as broadcasts, multicasts, and unknown unicasts. In an ELAN, the BUS is responsible for
servicing these traffic types by accepting broadcast, multicast, and unknown unicast packets
from the LECs via dedicated point-to-point connections, and forwarding the packets to all of
the members of the ELAN using a single point-to-multipoint connection. (Unknown unicast
packets are packets that the sending station broadcasts because it does not yet know the ATM
address for the packet’s destination MAC address. There may be more than one instance of an
active BUS per ELAN. Using ForeThought 5.0.x and greater, each BUS must be a colocated BUS
(also referred to as an intelligent BUS or a LES/BUS pair), which allows the BUS to use the
LES’s registration table to direct unicast traffic.

3.3 Emulated LAN Operation

This section describes the operation of an ELAN and its components from the point of view of
a LEC. The operation of an ELAN may be divided into three phases:
1. Initialization
2. Registration and Address Resolution
3. Data Transfer
ELAN components communicate with each other using ATM connections. LECs maintain
separate connections for traffic control functions and data transfer. The following connection
types are used by the LEC when operating in an ELAN:
• Configuration-Direct Connection: a temporary bidirectional point-to-point VCC set
up by the LEC to the LECS.
• Control-Direct Connection: a bidirectional point-to-point VCC set up by the LEC to
the LES. This connection must be maintained for the duration of the LEC’s partic-
ipation in the ELAN.
• Control-Distribute Connection: a unidirectional point-to-multipoint VCC set up by
the LES to the LEC. This connection must be maintained for the duration of the
LEC’s participation in the ELAN.
• Multicast-Send Connection: a bidirectional point-to-point VCC set up by the LEC to
the BUS for sending multicast data to the BUS. The LEC must attempt to maintain
this connection while participating in the ELAN.
• Multicast-Forward Connection: a unidirectional point-to-multipoint VCC set up
from the BUS to LECs participating in the ELAN. The LEC must attempt to main-
tain this connection while participating in the ELAN.
• Data-Direct Connection: a bidirectional point-to-point VCC set up between LECs
that want to exchange unicast data traffic, and torn down after 20 minutes
(default) of inactivity. Each LEC normally establishes many Data-Direct Connec-
tions.
For the following discussion, please refer to Figure 3.2.

LEC1 Ê CONFIGURATION - DIRECT LECS
Ë CONTROL - DIRECT
LES
Ì CONTROL - DISTRIBUTE
Í MULTICAST - SEND
BUS
Î MULTICAST - FORWARD
Ï DATA - DIRECT
engineering
LEC2
Configuring an
Emulated LAN
Figure 3.2 - ELAN Operation

3.3.1 Initialization
Upon initialization, LEC1 obtains its own ATM address via ILMI address registration. LEC1
obtains the address of the LECS in one of four ways: by querying the switch to which LEC1 is
connected via ILMI, by connecting to the “well-known” address defined by the ATM Forum’s
LANE standards (47.0079.00.000000.0000.0000.0000.00A03E000001.00), by using
PVC (0,17), or by using an address that is locally configured on LEC1.
Once it knows the location of the LECS, LEC1 establishes a configuration-direct connection Ê
to the LECS. When connected, the LECS provides LEC1 with the information necessary to
connect to the ELAN it wishes to join. This information includes such parameters as: the ATM
address of the ELAN’s LES, the type of LAN being emulated, the maximum packet size, and
the name of the ELAN (engineering, for example). This configuration information is con-
tained in a configuration file that must be built and maintained by the network administrator.
Detailed information about the LECS

NOTE configuration file may be found in Section 3.6.1.

3.3.2 Registration and Address Resolution

After obtaining the address of the LES, LEC1 establishes a control-direct connection Ë to the
LES.
When using DLE, this address is a single,

NOTE anycast address which allows the LEC to reach
one of the other DLE peer servers for its ELAN if
its local server goes down. This address is routed
via PNNI to the nearest active DLE peer server
for this ELAN.
If the LES is configured to perform ELAN access

NOTE control (see Section 3.5), upon receiving a request
from a LEC to join the ELAN, the LES sends a
message to the LECS to verify that the LEC is
allowed to join. If verification is received from
the LECS, then the LES gives the LEC permission
to join. If verification is not received from the
LECS, the LES rejects the join request and the
LEC is dropped.
The LES assigns LEC1 a unique identifier, and LEC1 registers its own MAC and ATM
addresses with the LES. (The LES maintains a table containing the MAC addresses and corre-
sponding ATM addresses of all members of the ELAN.) At this point, LEC1 has “joined” the
ELAN.
The LES then establishes a control-distribute connection Ì back to LEC1. Connections Ë and Ì
Configuring an
Emulated LAN
can now be used by LEC1 to send LAN Emulation ARP (LE_ARP) requests to the LES, and
receive replies.
LEC1 now sends an LE_ARP request to the LES to get the ATM address of the BUS corre-
sponding to the broadcast MAC address (FF-FF-FF-FF-FF-FF). The LEC then establishes a
multicast-send connection Í to the BUS. The BUS responds by setting up a multicast-forward
connection Î to the LEC.
At this point, the LEC is ready to transfer data.

3.3.3 Data Transfer

When LEC1 receives a network-layer packet from a higher layer protocol to transmit to some
destination MAC address (for example, LEC2), LEC1 initially does not know the correspond-
ing ATM address of the destination. Consequently, LEC1 transmits an LE_ARP request to the
LES.
The example shown in Figure 3.2 assumes that

NOTE LEC2 has already registered with the LES, and
that connections similar to those described for
LEC1 already exist.
While waiting for the LES to respond, LEC1 forwards the packet to the BUS. The BUS broad-
casts the packet to all LECs on the ELAN. This is done to avoid data loss, and to minimize con-
nection set-up latency (due to the LE_ARP process) that may not be acceptable to some
network protocols.
If the LE_ARP response is received, LEC1 establishes a data-direct connection Ï to the destina-
tion address of LEC2. This path will be used for subsequent data transfers. Before LEC1 begins
to use this connection, it first sends a “flush” packet via the BUS to the destination, LEC2.
When LEC2 acknowledges receipt of this packet, signifying that the BUS path is empty, only
then does LEC1 begin to use the data-direct connection Ï for data transfer. This process
ensures that the network protocol’s frames arrive in the proper order.
If no response is received to the LE_ARP, LEC1 continues to send data via the BUS, while con-
tinuing to LE_ARP until a response is received and a data-direct connection to LEC2 is estab-
lished.
If LEC1 already has a data-direct connection to a MAC address it wishes to reach, it need not
go through the LE_ARP process again. Instead, it continues to use the current connection. This
is possible because each LEC maintains a cache of MAC address to ATM address mappings
that it receives in response to the LE_ARPs it has sent. Entries in this cache are “aged” out over
a period of time. Data-direct connections are also cleared if they remain inactive for a period of
time.

3.4 Distributed LAN Emulation

Distributed LAN Emulation (DLE) allows the LES and BUS functions that are provided to
each ELAN to be distributed among multiple, interconnected server platforms. In this way,
DLE provides these ELANs with resiliency and scalability.
To understand DLE operation, it is useful to compare DLE to the current LANE service model,
which uses a single LES and BUS for each ELAN. This section first describes a simple example
of the single server model and then gives a detailed overview of the DLE model.
3.4.1 Single Server LANE Services Model

Figure 3.3 shows the topology of a single server supporting an ELAN. In this example, the
LECs are hosts that are using IP, and the LES and BUS are running on the same switch. Three
LANE LECs are all registered in the same ELAN called Eng, and each is, therefore, connected
to a LES and to a BUS for that ELAN.
Eng
LES/BUS 1
Eng LEC 1 Eng LEC 2 Eng LEC 3
Figure 3.3 - Single Server LANE Services Model
Configuring an
Emulated LAN

3.4.1.1 Using a Single Server

When LEC 1 wants to contact LEC 3, several messages are exchanged. First, LEC 1 attempts to
learn the MAC address of LEC 3 by broadcasting an IP-ARP request with LEC 3’s IP address.
As Figure 3.4 shows, this ARP request is sent in two steps: Ê as a point-to-point message from
LEC 1 to the LANE BUS, then Ë as a point-to-multipoint message from the BUS to all of the
LECs registered in the ELAN.
Eng
LES/BUS 1
Ê
Ë
Figure 3.4 - Broadcast IP-ARP Request
When LEC 3 receives the IP ARP request, it recognizes that it is the intended destination, and,
therefore, attempts to send an IP ARP response to LEC 1 (whose MAC address was supplied
in the ARP request packet).
As shown in Figure 3.5, the delivery of the ARP response is a three-step process: Ì LEC 3 sends
an LE-ARP query to the LES, asking for the ATM address that corresponds to LEC 1’s MAC
address; Í the LES sends an LE-ARP response to LEC 3; and Î LEC 3 establishes a circuit to
LEC 1’s ATM address.
Eng Í
LES/BUS 1
Î Ì
Figure 3.5 - IP ARP Response Handling

3.4.1.2 Limitations of a Single Server

Because the there is only one LES/BUS supporting the ELAN, the following limitations exist:
• The number of LECs in a single ELAN is limited by the number of virtual circuits
that the single LES/BUS can establish through their platform’s ATM port. This
usually limits the ELAN to about 500 LECs.
• Clusters of LECs that are geographically separated from the LES/BUS may have
poor throughput, even when connecting to each other, because address queries
and broadcasts may traverse slow wide-area links.
• A failure of the LES or BUS brings down the ELAN.
Configuring an
Emulated LAN

3.4.2 Distributed LAN Emulation Model

To address the limitations of the single server model, DLE distributes the LANE services load
among a mesh of LES/BUS DLE peer servers, as shown in Figure 3.6.
Eng Eng Eng

LES/BUS 1 LES/BUS 2 LES/BUS 3
Eng LEC 1 Eng LEC 2 Eng LEC 3 Eng LEC 4 Eng LEC 5 Eng LEC 6 Eng LEC 7 Eng LEC 8 Eng LEC 9
Figure 3.6 - Distributed LAN Emulation Model
Each DLE peer server actually maintains two sets of connections: one is a point-to-multipoint
connection to each of its peers for broadcasting multicast data and flooding control information,
and the other includes individual point-to-point connections to each peer for directed control
traffic.
Each DLE peer server that supports the ELAN is responsible for registering and giving reports
about the LECs that are attached to it directly. Each DLE peer server propagates this informa-
tion to both its locally attached LECs and its peers.
Each device running a DLE peer server must use

NOTE ForeThought 5.0 or greater; however, the DLE
peer servers support clients and attached
switches using ForeThought 4.0 and 4.1, and
third-party devices that are ATM Forum LANE
1.0 compliant.

3.4.2.1 Using DLE

Figure 3.7 shows how a connection begins to be established through DLE peer servers. LEC 1
wants to communicate with LEC 9, which is in the same ELAN, but is locally attached to a dif-
ferent DLE peer server. First, Ê LEC 1 sends an IP ARP broadcast request to its local DLE BUS.
Then, Ë the BUS broadcasts the packet to both its locally attached LECs and its DLE peer
servers.
Eng Eng Eng

LES/BUS 1 Ë LES/BUS 2 LES/BUS 3
Ê
Figure 3.7 - IP ARP Broadcast from LEC 1 to LEC 9
Upon receiving the broadcast from the first DLE peer server, the peers re-distribute the packet
to their own locally attached LECs Ì, as shown in Figure 3.8, so the packet arrives its actual
destination at LEC 9.
Eng Eng Eng

Ì
Configuring an
LES/BUS 1 Ì LES/BUS 3
Emulated LAN
LES/BUS 2
Figure 3.8 - Re-distributing the Broadcast across DLE Peer Servers
The peers do not re-distribute the packet to other

NOTE peers; this would create a loop.

LEC 9 recognizes its IP address, and prepares an IP ARP response. As shown in Figure 3.9, it
then sends an LE-ARP request to its local LES Í, asking for the ATM address that matches LEC
1’s MAC address. Since LEC 9’s local LES does not have an entry for LEC 1, the local LES
passes the query along to all of its locally-attached proxy LECs (none are shown in this figure)
and all of its DLE peer servers Î.
Eng Eng Î Eng

Í
Figure 3.9 - LE-ARP for Unknown Host Sent to Proxies (not shown) and DLE Peer Servers
In Figure 3.10, the second DLE peer server is attached to two proxy LECs (LEC 4 and LEC 5).
When the DLE peer server receives the LE-ARP query, it cannot resolve the query, so the DLE
peer server re-distributes the query to its proxy LECs Ï (but not to its peer servers again, to
avoid a loop). Meanwhile, the first peer server has been able to resolve the LE-ARP for the
address of LEC 1 and has sent an LE-ARP response to the third server Ð.
Ð
Eng Eng Eng
LES/BUS 1 Ï LES/BUS 2 LES/BUS 3
Figure 3.10 - LE-ARP Query Answered by One DLE Peer Server and Re-distributed by Another

When the third DLE peer server receives the LE-ARP response, it passes it directly to LEC 9
(which sent the original query) Ñ. The third DLE peer server also caches the registration infor-
mation for LEC 1 so that other local LECs do not have to go through the entire process again.
However, this cache ages out over time. LEC 9 can now open a connection to LEC 1, and send
its IP ARP response Ò, as shown in Figure 3.11.
Eng Eng Eng

Ñ
Figure 3.11 - LE-ARP Response Delivered and LEC 9 Contacts LEC 1
3.4.2.2 Advantages of DLE

As mentioned earlier, using DLE provides solutions to the problems of using a single server.
3.4.2.2.1 Load Sharing

DLE peer servers distribute the circuit and processing load. The number of LANE LECs is no
longer limited by the number of circuits one LES/BUS platform can maintain, since many
platforms can support a single ELAN. Also, more VCs are available for use by other applica-
tions.
Configuring an
Emulated LAN
3.4.2.2.2 Improved Performance for Remote LECs
With DLE, broadcast delivery and LE-ARP resolution across peer servers can take a little
longer than if all LECs were connected to a single server, since extra processing steps and
transmissions are needed. However, ELANs with groups of LECs in different locations can be
designed for higher performance by providing a DLE peer server with each group. Broadcasts
and address resolution within each group will improve.
3.4.2.2.3 Fault Tolerance

Perhaps the most important advantage of DLE is fault tolerance. In a single server ELAN, the
server can be a single point of failure. If the server fails, endstations in the ELAN are unable to
discover each other through broadcast queries and unable to resolve MAC addresses into
ATM addresses. Increased network reliability, therefore, requires that ELANs have backups
for LES and BUS functions. To illustrate this point, the single server model is again discussed.

3.4.2.2.3.1 Single Server ELAN

Figure 3.12 shows a single server ELAN composed of nine LECs attached to three different
switches. The LECS and the LES/BUS are attached to a host connected to a single switch. The
process for LEC 1 to connect to the LANE services takes several steps:
1. LEC 1 asks the signalling software on its switch to open a connection to the “well-
known” LECS address. (Other addressing methods could also be used).
2. The signalling software knows that this address is attached to port N (the port on
which the host resides on the switch), and opens a circuit between LEC 1 and port
N.
3. LEC 1 sends a message to the LECS, asking for the address of the LES for LEC 1’s
ELAN. The LECS responds with the ATM address of the LES, and LEC 1 estab-
lishes a circuit to the LES and then the BUS.
ELAN Eng
LES/BUS
Ì
LECS
Ë
Switch 1 Switch 2 Switch 3
Figure 3.12 - ELAN with a Single Server and Multiple Switches Connecting to Services
In Figure 3.13, when LEC 7 goes through the same process, it is slightly more complicated.
When LEC 7 asks its local switch’s signalling software to establish a circuit to the LECS Ê, the
local switch must use inter-switch link information (IISP or PNNI tables) to establish a cross-
switch circuit to the LECS. Once this circuit is established, however, the process is identical.

ELAN Eng
LES/BUS 1
Î
Ì Ë
LECS
Í
Figure 3.13 - ELAN with Single Server and Remote Connection to Server
Figure 3.14 shows the ELAN in operation after three LECs have gone through the registration
process.
ELAN Eng
LES/BUS 1
LECS
Configuring an
Emulated LAN
Figure 3.14 - ELAN with Single Server in Operation
If the single server in Figure 3.14 goes down, then the entire ELAN goes down. At this point,
the administrator must intervene and reconfigure the ELAN.

3.4.2.2.3.2 DLE ELAN

As noted previously, having a single server supporting an ELAN has a potential problem
because the server can be a single point of failure. However, DLE can address this problem. By
attaching the ELAN LECs to multiple DLE peer servers which communicate with each other
as described earlier, the number of LECs affected by a server failure is reduced, and a backup
server is provided for affected LECs to use. Figure 3.15 shows the configuration of such an
ELAN as three stations register.
ELAN Eng ELAN Eng

LES/BUS 1 LES/BUS 2
LECS Ë
Ë Ë
Ë
Ê Ê
Ê
Figure 3.15 - Registrations on an ELAN with Multiple Servers
LECs 1, 4, and 7 are directed by their switches to the LECS. The result is shown in Figure 3.16.
The connection between the two servers carries

NOTE broadcasts and LE-ARP traffic as described
earlier.

ELAN Eng ELAN Eng

LES/BUS 1 LES/BUS 2
LECS
Figure 3.16 - ELAN with Multiple Servers in Operation
This ELAN may experience significant performance improvements for the reasons described
earlier. Even if the actual performance is similar to using a single server in a particular net-
work, a great advantage is gained through its fault-tolerance if one of the servers fails as
depicted in Figure 3.17.
ELAN Eng ELAN Eng

LES/BUS 1 LES/BUS 2
Configuring an
Ë
Emulated LAN
LECS
Ë Ë
Ê Ê
Figure 3.17 - Failure of One ELAN Server and the Recovery Process

The failure and recovery process occurs as follows:

1. Eng LES/BUS 1 has lost power. All circuits connected to it are torn down. Low-
level signalling traffic (e.g., SSCOP messages) stop, and Switch 1 removes the
address of Eng LES/BUS 1 from its link tables.
2. LECs 1 and 4 had been connected to Switch 1. They detect that their connections to
Eng LES/BUS 1 have been torn down; user intervention is not necessary.
3. LECs 1 and 4 follow LANE 1.0 protocols to locate an LECS again to find the address
of their ELAN’s LES. In this example, they again connect to the LECS.
4. The LECS reports to LECs 1 and 4 that their ELAN server is at ATM address N1.
This address is used by every peer LES supporting the ELAN; both Eng LES 1 and
Eng LES 2 in this example.
5. LEC 1 sends a request to Switch 1 to establish a connection to address N1. Switch
1 no longer believes it has a direct connection to N1, and instead uses PNNI to
establish a circuit through Switches 2 and 3 to Eng LES 2.
6. LEC 4 sends a request to Switch 2 to establish a connection to N1. Switch 2 may
have learned from Switch 1 that it no longer offers a connection to N1, or Switch 2
may attempt a route through Switch 1 and be “bounced back” through ATM
Forum PNNI crankback. Either way, Switch 2 finally routes the connection through
Switch 3 to Eng LES 2.
This recovery process occurs quickly -- clients typically recover at a rate of 100 clients per
minute -- and the result is a re-configured ELAN as shown in Figure 3.18.
ELAN A ELAN A
LES/BUS 1 LES/BUS 2
LECS
Figure 3.18 - ELAN Re-established Using the Second Server

3.5 ELAN Access Control

Basic ATM Forum LAN Emulation Servers do not guard against unauthorized users learning
an ELAN’s LES address and then joining the ELAN. However, a method of authorization
checking is available. After a LEC obtains the address of its LES, the LEC sends a request to
the LES to join the ELAN. If the LES has ELAN access control enabled, it sends a message to
the LECS to verify that the LEC is allowed to join. If verification is received from the LECS,
then the LES gives the LEC permission to join. If verification is not received from the LECS,
the LES rejects the join request and the LEC is dropped.
Using this feature, an authorization check is also performed each time the LECS reloads the
LECS configuration file. (The LECS periodically checks whether its configuration file has been
modified, and, if it has, the file is re-read. The length of this period, in seconds, is defined by
the Reload_Period key.) If the file has changed to disallow some clients that were previ-
ously allowed, those clients will be dropped from the ELAN.
ELAN access control also works with a third-
NOTE party LECS. The LES revalidates the client every
600 seconds since the third-party LECS will not
contact the LES with configuration changes.
You can enable ELAN access control when you are creating the LES. When you use the
services les new command, specify the -secure option as enabled. This indicates you
want to activate a secure LES/BUS pair and a secure well-known LECS address. (ELAN access
control is disabled by default.)
However, if you are using an LECS address that is different than the well-known address, then
you must also type the full LECS ATM address to be used using the [-lecs] <NSAP
Address> option.
If you want to disable ELAN access control, or if you want to enable ELAN access control at a
Configuring an
Emulated LAN
later time after the LES has been created, you can do this by entering the services les new
command and specifying a value for the -secure option. See Chapter 2 in Part 3 of the AMI
Configuration Commands Reference Manual for more information. By using this command, you
do not have to delete and recreate the LES.

3.6 Configuring an ELAN
When configuring LANE services in your

NOTE network, it is highly recommended that you use
a switch that has a 32MB SCP or greater to run
those services.
There are different instructions for configuring an ELAN, depending on how your network is
currently configured. Please read the following list to determine which set of instructions to
use.
• If you had previously configured LANE, you want to upgrade some or all of the
clients to ForeThought 6.x, and you want to upgrade all the equipment that is run-
ning services to ForeThought 6.x using DLE, use the instructions found in
Section 3.7.
• If you had previously configured LANE, you want to leave the clients running
ForeThought 4.1.x, and you want to upgrade all the equipment that is running ser-
vices to ForeThought 6.x without using DLE, use the instructions found in
Section 3.8.
• If you are configuring LANE for the first time and all of your equipment is run-
ning ForeThought 6.x, use the instructions that follow here in Section 3.6.
To configure an ELAN on a switch, you must log into AMI on a switch running ForeThought
6.1.x and use the commands found under services lane lecs, services lane les, and
interfaces lec.
More information about each of these commands

NOTE may be found in Chapter 2 in Part 3 of the AMI
Configuration Commands Reference Manual.
There are three major steps that the system administrator should follow in order to configure
and maintain ELANs:
1. Configure an LECS configuration database file.
2. Start the LAN Emulation Services (LECS and LES/ BUS).
3. Start the LEC(s) and join an ELAN.

Steps 1 may be performed using a text editor on

NOTE any system. However, the resulting file can be
used only on systems running under Solaris 2.5,
2.5.1, or 2.6, or on a FORE switch running
ForeThought 4.0 or greater or a TNX switch
running ForeThought 5.1 or greater.
The remainder of this section gives a practical example of configuring and administering an
ELAN.
Configuring an
Emulated LAN

3.6.1 Configuring an LECS Configuration Database File

The LECS uses a text configuration file to contain the configuration information needed by
LECs that wish to participate in an ELAN. The LECS configuration file may be built and
edited using a text editor such as vi or emacs.
ForeThought VLAN Manager, a product available

NOTE separately from FORE Systems, greatly
simplifies the administration of ELANs. Its
graphical user interface controls the content of
the LECS configuration file transparently to the
user. For more information, please refer to the
ForeView VLAN Manager User’s Manual.
3.6.1.1 Before You Begin

Before building or modifying the LECS configuration file, you should first determine the
topology of the ELAN or ELANs that you want to administer. You must supply the following
information when building or editing the LECS configuration file:
• Provide the name of each ELAN (engineering, marketing, etc.).
• Provide the LAN type (Ethernet or Token Ring) for each ELAN.
• Provide the MTU size for each ELAN.
• Provide the ATM address of the LES for each ELAN. If you are using DLE, this
address must be the anycast address for the DLE peer servers in each ELAN. Be
sure to choose a distinct anycast address for each ELAN in the network. It must be
unique within the first 19 bytes.
• Provide the address of each LEC that may participate in each ELAN.
• Provide the MPOA control parameters if you wish to run MPOA.
• If you wish LECs to use a default ELAN, the default LES information must also be
included.
• Provide various other configurable parameters.
CAUTION Do not attempt to edit an existing functional

LECS configuration file without first making a
backup copy of the file. Incorrect modification of
the configuration file could result in loss of
communication between one or more members
of a defined ELAN.

You may make changes to the LECS

NOTE configuration file while the LECS process is
running. The configuration file is reread
periodically by the LECS process (the default
period is ten minutes). Consequently, any
changes that you make to the configuration file
are not recognized until the file is reread.
3.6.1.2 LECS Configuration File Syntax

Each line that you enter in the configuration file takes the general form:
[[group].]key : value
The group field may represent:

• ELANs (by name) - ELAN names are case-sensitive, and may not exceed 32 char-
acters in length
• clients ATM or MAC addresses
• miscellaneous LECS control information specified by using a group name of
LECS
The key field is used to denote an individual parameter within a group.
The value field contains the value assigned to the key.
Omitting the group implies that the key and value apply to all groups in the configuration
file. Leading and trailing spaces, as well as spaces on either side of the “:”, are ignored.
For example, to specify a maximum frame size of 1516 bytes for the ELAN named
engineering, enter the following:
Configuring an
Emulated LAN
engineering.Maximum_Frame_Size : 1516
Similarly, to specify a default maximum frame size of 1516 bytes for all ELANs defined in a
given configuration file, enter the following:
.Maximum_Frame_Size : 1516

Table 3.1 defines the various key parameters that may be entered in the configuration file. The
acceptable range of values and the default value for each parameter is also given.
Table 3.1 - LECS Configuration File Parameters
Parameter Definition
.LAN_Type: Ethernet/IEEE 802.3 Identifies the type of ELAN, either Ethernet/IEEE 802.3
or IEEE 802.5. The default is Ethernet/IEEE 802.3.
.Maximum_Frame_Size: 1516 Specifies the length (in number of bytes) of the largest
frame. Selections are: 1516, 1580, 4544, 9234, and 18190.
The default is 1516 for Ethernet and 4544 for Token Ring.
.Control_TimeOut: 120 Specifies the timing out of request/response control
frame interactions, in seconds. The minimum is 10 sec-
onds and the maximum is 300 seconds. The default is
120 seconds.
.Maximum_Unknown_Frame_Count: 1 Limits the number of unicast frames sent to the BUS. The
minimum is 1 frame per 60 seconds and the maximum is
10 frames per second. The default is 1 frame per second.
.Maximum_Unknown_Frame_Time: 1 Limits the number of unicast frames sent to the BUS in
the specified number of seconds. The default is 1 second.
.VCC_TimeOut_Period: 1200 Specifies the length of time that an idle data connection
remains open before being closed. The default value is
1200 seconds.
.Maximum_Retry_Count: 1 Limits the number of LE_ARP retransmission requests.
The minimum is 0 and the maximum is 2. The default is
1.
.Aging_Time: 300 Specifies the period that LE_ARP cache table entries
remain valid, in seconds. The minimum is 10 and the
maximum is 300. The default value is 300 seconds.
.Forward_Delay_Time: 15 Specifies the timing out of non-local ARP cache entries
in seconds. The minimum is 4 and the maximum is 30.
The default value is 15 seconds.
.Expected_LE_ARP_Response_Time: 1 Specifies the maximum time a LEC expects an LE_ARP
request/response will take, in seconds. The minimum is
1 and the maximum is 30. The default value is 1 second.
.Flush_TimeOut: 4 Specifies the maximum time a LEC expects an
LE_FLUSH request/response will take, in seconds. The
minimum is 1 and the maximum is 4. The default value
is 4 seconds.

Table 3.1 - LECS Configuration File Parameters (Continued)
.Path_Switching_Delay: 6 Specifies the minimum time between switching BUS and
data paths, in seconds. The minimum is 1 and the maxi-
mum is 8. The default value is 6 seconds.
.Local_Segment_ID The segment ID of the emulated LAN for an IEEE 802.5
source routing bridge.
.ELAN_Identifier: auto Specifies a non-zero ELAN identifier. The default is
auto, which signals the LES to generate an ELAN identi-
fier itself.
.Multicast_Send_VCC_Type: Specifies the multicast send mode, either Best Effort,
Best Effort Variable, or Constant. The default is Best Effort.
.Connection_Complete_Timer: 4 Specifies the time period in which data or READY_IND
is expected, in seconds. The minimum is 1 and the maxi-
mum is 10. The default is 4 seconds.
.ShortCut_Protocols: IP Specifies the set of protocols on which to perform flow
detection. Also, specifies the set of protocols for which
MPOA resolution is supported. The default is {}.
Currently, the only supported value is IP.
.ShortCut_Threshold: 10/1 Specifies the number of frames per second that a LANE/
MPOA client (LEC/MPC) forwards to the same destina-
tion via the default forwarding path before which it
should begin using a shortcut. The minimum number of
frames is 1 frame and the maximum is 65,535 frames.
The minimum rate of speed at which the frames are for-
warded is 1 second and the maximum is 60 seconds. The
Configuring an
Emulated LAN
default is 10 frames per second.
.Resolution_Initial_Retry_Time: 5 Specifies the initial retry time interval after which a
LEC/MPC may send another MPOA Resolution
Request if an MPOA Resolution Reply has not been
received for the initial request. The minimum is 1
second. The maximum is 300 seconds. The default is 5
seconds.
.Resolution_Maximum_Retry_Time: 40 Specifies the maximum retry time interval after which a
LEC/MPC assumes an MPOA Resolution Request has
failed. The minimum is 10 seconds. The maximum is 300
seconds. The default is 40 seconds.

.Resolution_Hold_Down_Time: 160 Specifies the minimum amount of time to wait before re-
initiating an MPOA Resolution Request after a failed
resolution attempt. This value is usually greater than the
Resolution_Maximum_Retry_Time. The minimum is 30
seconds. The maximum is 1,200 seconds. The default is
160 seconds.
.MPOA_KeepAlive_Time: 10 Specifies how often an MPS must send MPOA Keep-
Alive messages to all LEC/MPCs for which it has cre-
ated and is maintaining Egress Cache Entries. The mini-
mum is 1 second. The maximum is 300 seconds. The
default is 10 seconds.
.MPOA_KeepAlive_Lifetime: 35 Specifies the length of time that a Keep-Alive message is
considered valid. It is recommended that this value be at
least twice the value of the MPOA_KeepAlive_Time.
The minimum is 3 seconds. The maximum is 1,000
seconds. The default is 35 seconds.
.Resolution_GiveUp_Time: 40 Specifies the minimum amount of time to wait before an
MPS should give up on a pending resolution request.
The minimum is 5 seconds. The maximum is 300 sec-
onds. The default is 40 seconds.
.Resolution_Default_Holding_Time: Specifies the default holding time for use in NHRP Reso-
1200 lution Replies. The minimum is 60 seconds. The maxi-
mum is 7,200 seconds. The default is 1,200 seconds.
.IP_Flows: <ip-flow> Specifies a set of IP flows that are defined by the param-
{,<ip-flow>} eters listed below.
<ip-flow> Specifies the IP information that is defined by the
parameters listed below. The format is <src-dst>
[<proto>] <thresholds> <qos>.
<src-dst> Specifies the source and destination addresses of an IP
flow. The format is <ip-addr> <ip-addr> with the
source address first and the destination address second.
<ip-addr> Specifies the format for the IP addresses <ip-addr>
used above. The format is <dotted-addr>/<prefix-
leng>.

<dotted-addr> Specifies the format for the IP addresses <dotted-
addr> used above. The format is <octet>.<octet>.
<octet>.<octet>. The range of each octet is from 0 to
255.
<prefix-leng> Specifies the prefix length to indicates how much of a
given IP address is significant. The range is 0 through 32.
0 means any IP address can be matched and 32 means
the entire IP address is significant.
<proto> Optionally specifies for which protocol this flow applies
(ICMP, IGMP, TCP, or UDP). The format is one of: ICMP,
IGMP, TCP <port> <port>, or UDP <port> <port>. If
no protocol is specified, then this flow applies to all pro-
tocols.
<port> Specifies the source and destination port number,
respectively, to be used with the <proto> parameter. 0
indicates a match for any port number. The range is 0
through 65535.
<thresholds> Specifies a traffic threshold that must be reached before
a shortcut VCC is established for a flow. Also, specifies a
threshold below which a shortcut VCC is no longer con-
sidered valid, is torn down, and the default forwarding
path is used. Both thresholds are expressed in frames per
second. The format is <threshold> <threshold>
<threshold> with the set-up threshold first, the tear-
down threshold second, and the set-up threshold for the
Configuring an
Emulated LAN
default forwarding path third.
<threshold> Specifies the format for entering the threshold data as
described for <thresholds>. The format is <frame-
count>/<frame-time>. For set-up, a zero count with
a non-zero time means a shortcut VCC should be estab-
lished upon the first frame and a non-zero count with a
zero time means a shortcut VCC is never established.
For tear-down, a zero count with a non-zero time means
a shortcut VCC is never torn down.
<frame-count> Specifies the format for the number of frames for the set-
up and tear-down thresholds. The range is 0 through
65535.

<frame-time> Specifies the format for the number of seconds for the
set-up and tear-down thresholds. The range is 0 through
65535.
<qos> Specifies the parameters used in signalling a connection
for the specified flow. The format is <qos-flags>
<qos-class>. Currently, <qos-flags> is unused and
its value should be 0.
<qos-class> Specifies the Quality of Service (QoS) class for this flow.
Shared specifies a UBR connection will be shared
among all flows to the same destination. UBR specifies a
non-shared UBR connection. The format is UBR <rate>.
CBR specifies a CBR connection. The format is CBR
<rate>. VBR specifies a VBR connection. The format is
VBR <rate> <rate> <burst> (for CLP=0+1 cells)
<rate> <burst> (for CLP=0 cells). NRT-VBR specifies
a non-real time VBR connection. The format is NRT-VBR
<rate> <rate> <burst> (for CLP=0+1 cells) <rate>
<burst> (for CLP=0 cells). ABR specifies an ABR con-
nection. There is no format currently defined for ABR.
<rate> Specifies the rate for the QoS class specified. For UBR
and CBR, this is the Peak Cell Rate (PCR) in cells per sec-
ond. For VBR and NRT-VBR, this is the PCR for cells that
are CLP=0+1 in cells per second, the Sustainable Cell
Rate (SCR) for cells that are CLP=0+1 in cells per second,
and the SCR for cells that are CLP=0 in cells per second.
<burst> Specifies the Maximum Burst Size (MBS) in cells for the
QoS class specified (VBR and NRT-VBR only). This is the
MBS for cells that are CLP=0+1, in cells, and the MBS for
cells that are CLP=0, in cells.
Lines beginning with # may be inserted if you wish to include comments or to improve the
clarity of the presentation when the file is viewed or printed. These lines are ignored when the
file is read. Lines may be continued by escaping the end-of-line with a backslash “\” (do not
enter the quote marks).

3.6.1.3 Defining an ELAN

Each ELAN is defined by an address statement whose value denotes the ATM address of the
ELAN’s LES. For example:
engineering.Address: C5.0005.80.ffe100.0000.f21a.01b9.0020480605b2.00
To configure DLE for an ELAN, use the anycast

NOTE address in this statement. Be sure to use a
distinct anycast address for each ELAN in the
network. It must be unique within the first 19
bytes.
In addition, you may instruct a given ELAN to override any of the default values. For exam-
ple, the engineering ELAN could override a Maximum_Frame_Size setting of 1516; thus:
engineering.Maximum_Frame_Size: 4544
If you want to control which clients may or may not join a given ELAN, two additional keys,
Accept and Reject, whose values are comma-separated lists of matching elements, may be
used.
The Accept values must be supplied if you are

NOTE going to enable ELAN access control.
These values may be:
Configuring an
Emulated LAN
a MAC address,
engineering.Accept: 0020480605b2 , 002048080011 , 0020481020ef
an ATM address and equal-length bit mask,
engineering.Accept: 47.0005.80.FFE100.0000.0000.0000.002048000000.00\
FF.FFFF.FF.FFFFFF.0000.0000.0000.FFFFFF000000.00
or an ATM address containing “don’t-care” semi-octets denoted by an “X”:
marketing.Accept: 47.0005.80.FFE100.XXXX.XXXX.XXXX.002048XXXXXX.XX

The last two forms of ATM-address matching elements are functionally the same. The latter is
shorter but only allows for masks whose semi-octets are all ones or all zeros, while the former
allows for arbitrary masks. A prospective-client address is “captured” by an ELAN name if
the client’s address matches one of the Accept elements but not one of the Reject elements
(if present). Finally, an ELAN may be configured to accept any client that wishes to join by
including the following statement:
default.Accept:XX.XXXX.XX.XXXXXX.XXXX.XXXX.XXXX.XXXXXXXXXXXX.XX
The order in which to apply the Accept and Reject rules is given by a Match.Ordering
group.key statement, whose value is a comma-separated list of ELAN names. For example:
Match.Ordering: default, engineering, marketing
The names of all ELANs that have Accept keys must be included in Match.Ordering.
The LE_CONFIGURE_REQUEST frame contains an ATM address and an optional MAC
address or route descriptor (which is always present for ELAN access control requests). The
Accept/Reject checking proceeds in two distinct phases: first, for the MAC address or route
descriptor, if present, and second, for the ATM address. So, even though a client might be
rejected by its MAC address, it can be accepted by its ATM address. Therefore, when configur-
ing the Accept and Reject rules, ensure that you write them either as explicit lists of only MAC
addresses or route descriptors, or only as ATM address matches.
3.6.1.4 Defining a Client

Clients need not be defined in the LECS configuration file. Typically, you would define a client
for the purpose of overriding one or more of the default configuration parameters for that par-
ticular client.
A client is defined by using its ATM address, MAC address, or route descriptor in the group
field, and perhaps giving the name of its ELAN as the value of the LAN_Name key. For
example:
47.0005.80.FFE100.0000.0000.0000.002048222222.22.LAN_Name:engineering
002048ABCDEF.LAN_Name: marketing
012F.LAN_Name: publications
Configuration parameter overrides can also be given on a per-client basis. For example, the
following statements override the default VCC_TimeOut_Period and Aging_Time configu-
ration parameters for a client whose MAC address is 002048080011 on the engineering
ELAN:
002048080011.LAN_Name:engineering
002048080011.VCC_TimeOut_Period:1200
002048080011.Aging_Time: 30

3.6.1.5 LECS Control Parameters

Specifying values for keys in the LECS group provides control over the operation of the LECS
process.
If you change the values of the LECS control

NOTE parameters while the LECS process is running,
the new values do not take effect until the LECS
process is stopped, and then restarted.
When a client contacts the LECS, the connections established are known as Configuration
Direct VCCs. To override the default value of the VCC_TimeOut_Period key (the number of
seconds before an idle Configuration Direct VCC is automatically closed by the LECS), enter a
statement similar to the following:
LECS.VCC_TimeOut_Period: 1200
The LECS periodically checks whether its configuration file has been modified, and, if it has,
the file is reread. The length of this period, in seconds, is given by the Reload_Period key:
LECS.Reload_Period: 600
The Permanent_Circuits key holds a comma-separated list of VPI.VCI pairs denoting the
local ends of 0.17 PVCs on which the LECS should listen. For example:
LECS.Permanent_Circuits: 0.42, 0.112
The LECS can provide the client with a fourteen-bit pattern to permute the MAC-address gen-
eration algorithm. This bit pattern is specified with the MAC_Address_Base key.
Configuring an
Emulated LAN
LECS.MAC_Address_Base: 38fe

3.6.1.6 LECS MPOA Parameters

MPOA requests use similar, and in many cases, the same database keys as LANE requests.
However, there are some MPOA-specific keys that can be specified. LEC/MPC parameters
can be specified for shortcuts. For example, the following group of parameters indicates that
shortcuts should be established for IP flows, but only after a threshold of 10 frames per second
is surpassed.
.ShortCut_Protocols: IP
.ShortCut_Threshold: 10/1
LEC/MPC parameters can also be specified for resolution requests. When a LEC/MPC sends
an MPOA Resolution Request, it sets a timer to a Resolution_Initial_Retry_Time. If an MPOA
Resolution Reply is not received in that amount of time, a retry may be sent. Each time a retry
is sent, the timer is set to the Resolution_Initial_Retry_Time value * a retry multiplier. If the
value exceeds the Resolution_Maximum_Retry_Time value, the LEC/MPC assumes the
request has failed. A new request may not be sent until the Resolution_Hold_Down_Time has
been surpassed.
For example, the following parameters indicate that an initial MPOA Resolution Request
should be retried after 5 seconds, backing off to a maximum retry time of 40 seconds, and then
the MPOA Resolution Request process is re-initialized after 160 seconds.
.Resolution_Initial_Retry_Time: 5
.Resolution_Maximum_Retry_Time: 40
.Resolution_Hold_Down_Time: 160
MPOA server (MPS) parameters can also be specified. For example, the following parameters
indicate that the MPS must send Keep-Alive messages every 10 seconds. Each of these mes-
sages is valid for 35 seconds. The MPS must wait 40 seconds before giving up on a pending
resolution request, and should use 1200 seconds as the default holding time in NHRP Resolu-
tion Replies.
.MPOA_KeepAlive_Time: 10
.MPOA_KeepAlive_Lifetime: 35
.Resolution_GiveUp_Time: 40
.Resolution_Default_Holding_Time: 1200
Parameters can also be specified for flow descriptors which determine whether and when to
trigger creation of shortcut circuits. The LECS sends the LEC/MPC this set of parameters.
These parameters consist of the following elements in the following order: a source/destina-
tion specifier, flow establishment thresholds, and a QoS descriptor.

For example, you could specify that telnet traffic to the Class C 202.19.88.0 subnet should be
sent on a UBR VCC with a peak rate of 10,000 cells per second, but only after the traffic on that
connection exceeds 20 frames per second. If the VCC is idle for more than 10 minutes, then the
shortcut should be torn down.
.IP_Flows:0.0.0.0/0 202.19.88.0/24 \
TCP 0 23 \
20/1 1/600 0/1 \
0x0 UBR 10000
The parameters in the IP flow example are defined as follows:

0.0.0.0/0 The source address/prefix length.
202.19.88.0/24 The destination address/prefix length.
TCP 0 23 The protocol to which this flow applies, and the
source and destination ports, respectively.
20/1 1/600 0/1 The set-up threshold, the tear-down threshold, and
the set-up threshold for the default forwarding path,
respectively.
0x0 UBR 10000 The QoS flag (currently unused and should be set to
0), the QoS class, and the PCR for the specified QoS
class, respectively.
Configuring an
Emulated LAN

3.6.2 Sample LECS Configuration File
CAUTION Do not attempt to edit an existing functional

LECS configuration file without first making a
backup copy of the file. Incorrect modification of
the configuration file could result in loss of
communication between one or more members
of a defined ELAN.
For a detailed discussion of how to configure an

NOTE LECS configuration file similar to the one given
in this section, please refer to Section 3.6.1.
The sample LECS configuration file shown at the end of this section in Figure 3.19 and
Figure 3.20 defines three ELANs:
• default
• engineering
• marketing
The Match.Ordering statement specifies the ELAN names in the order that prospective cli-
ents will attempt to match. The default configuration parameters are shown with their default
values. These values apply to all ELANs in this configuration file, unless overridden for a par-
ticular ELAN or client.
ELAN default is configured to accept any client that wishes to join. The ATM address of the
default LES is listed in the default.Address statement. If DLE is being configured for
ELAN default, then this address must be the anycast address that allows the clients to reach
any of the DLE peer servers for ELAN default.
Be sure to choose a distinct anycast address for

NOTE each ELAN in the network. It must be unique
within the first 19 bytes.
ELAN engineering has overridden the default Maximum_Frame_Size with a new size of
4544 bytes. Consequently, this frame size applies only to traffic on the engineering ELAN.
The default and marketing ELANs continue to use the default frame size of 1516 bytes.
Two LECs, whose MAC addresses are 002048080011 and 0020481020ef, are identified as
acceptable clients for the engineering and marketing ELANs.

#
# The search ordering of elan names
#
Match.Ordering: default, engineering, marketing
#
# the default configuration parameters
#
.Control_TimeOut: 120
.Maximum_Unknown_Frame_Count: 1
.Maximum_Unknown_Frame_Time: 1
.VCC_TimeOut_Period: 1200
.Maximum_Retry_Count: 1
.Aging_Time: 300
.Forward_Delay_Time: 15
.Expected_LE_ARP_Response_Time: 1
.Flush_TimeOut: 4
.Path_Switching_Delay: 6
.Multicast_Send_VCC_Type: Best Effort
.Connection_Complete_Timer: 4
.LAN_Type: Ethernet/IEEE 802.3
.Maximum_Frame_Size: 1516
.ShortCut_Protocols:IP
.ShortCut_Threshold:10/1
.Resolution_Initial_Retry_Time:5
.Resolution_Maximum_Retry_Time:40
.Resolution_Hold_Down_Time:160
.MPOA_KeepAlive_Time:10
Configuring an
Emulated LAN
.MPOA_KeepAlive_Lifetime:35
.Resolution_GiveUp_Time:40
.Resolution_Default_Holding_Time:1200
.IP_Flows: 0.0.0.0/0 202.19.88.0/24 \
TCP 0 23 \
20/1 1/600 0/1 \
0x0 UBR 10000
#
# Parameters for the default elan
#
default.Address: C5.0005.80.ffe100.0000.f21a.21b8.0097036324b2.25
default.Accept: XX.XXXX.XX.XXXXXX.XXXX.XXXX.XXXX.XXXXXXXXXXXX.XX
Figure 3.19 - Sample LECS Configuration File (Part One of Two)

#
# Parameters for elan: engineering
#
engineering.Address:C5.0005.80.ffe100.0000.f21a.01b9.0020480605b2.11
engineering.Accept: 002048080011 , 0020481020ef
engineering.Maximum_Frame_Size: 4544
#
# Parameters for elan: marketing
#
marketing.Address: C5.0005.80.ffe100.0000.f21a.01b9.0020480605b2.21
marketing.Accept: 002048080011 , 0020481020ef
#
Figure 3.20 - Sample LECS Configuration File (Part Two of Two)

3.6.3 Starting the LAN Emulation Services

LAN Emulation services include the LECS and the LES/BUS. Once the LECS configuration
database file has been configured, these services must be started so that they are available for
LECs to attempt to use. Using ForeThought 5.0 or greater, the LES/BUS services must run in
the same device.
LANE services running on a switch are LANE

NOTE 2.0 compliant, but LANE clients running on a
switch are LANE 1.0 compliant.
3.6.3.1 Starting the LECS

Once an LECS configuration file has been configured, you need to retrieve the LECS configu-
ration database file that you built elsewhere and put it on the switch that is going to run the
LECS.
1. Use the following AMI command to retrieve the file:
services lane lecs get [-url] <URL>
For example, you would enter something similar to the following:
services lane lecs get 198.29.22.46/lecs.cfg
2. After you have retrieved the LECS configuration database file, use the following
AMI command to start the LECS service on the switch:
services lane lecs new
Configuring an
Emulated LAN
Usage:
[[-index] <integer>] Index (default: 1)
[-atmaddress] <NSAP Selector> LECS ATM Address
[[-adminstatus] (up|down)] Admin State
[[-database] <text>] Database File Name
[[-defaultles] <NSAP Address>] Default LES ATM Address
[[-lesflag] (enabled|disabled)] Default LES Flag
[[-wkaflag] (none|atmforum|other)] WKA Flag
[[-wka] <NSAP Address>] Well Known Address

For example, to start the LECS service using the -db option and using the ATM
Forum’s well-known address you would enter something similar to the following:
services lecs new -atmaddress 0x0c -database lecs.cfg -wkaflag atmforum
If you want to use an address other than the

NOTE ATM Forum’s address, enter the -wka option
and the address. If you want to disable the well-
known address so that the LECS can only be
contacted by using the switch’s actual address
(with selector byte), use -wkaflag none.
3. Use the following AMI command to verify that the LECS has been started and is
running:
services lane lecs show
Index Admin State Oper State WKA Database

1 up up atmforum lecs.cfg
LECS: 0x47.0005.80.ffe100.0000.f21a.5208.0020481a5208.f1
WellKnown: 0xc5.0079.00.000000.00000000000000a03e000001.00
Default LES: 0x47.0005.80.ffe100.0000.f21a.5208.0020481a5208.f0
The OperStatus field shows up, meaning that the LECS is running. Now you
must start the DLE peer servers as described in the next section.
If you used an address other than the ATM

NOTE Forum’s address, the WKA field would show
other and that address would be displayed
below the entry. If you disabled the well-known
address the WKA field would show none.
If you used the -default option, then that LES

NOTE address would be displayed below this entry in a
field titled Default LES.

3.6.3.2 Starting the DLE LES/BUS Peer Servers

The LES and BUS services must be started for the ELAN. This example assumes you are using
DLE; therefore, you must enter a distinct anycast address (that is unique within the first 19
bytes) for the LECs to use to contact the LES, and the address of each of the DLE peer servers
so that this server can communicate with its peers.
1. To start the services, use the following AMI command on the switch that is going
to run one of the DLE peer servers:
services lane les new
Usage:
[-atmaddress] <NSAP Selector> LES ATM address
[-name] <text> ELAN Name
[[-adminstatus] (up|down)] Admin State (default: up)
[[-bus] <NSAP Address>] BUS ATM Address
[[-type] (ethernet|token-ring)] LAN Type (default: ethernet)
[[-mtu] <mtu>] MTU (default: 1516)
[[-lecs] <NSAP Address>] Secure LECS Address
[[-secure] (enabled|disabled)] Secure
[[-fwdtlvs] (enabled|disabled)] Forward TLVs
[[-elanid] <integer>] ELAN ID
[[-trnumber] <Hex Integer>] Token Ring Number
[[-fwdarp] (enabled|disabled)] Forward ARP
[[-anycast] <NSAP Address>] Anycast ATM Address
[[-peerlist] <text>] Peer ATM Addresses
Configuring an
Emulated LAN
services les new -atmaddress 90 -name engineering -anycast
c5.0005.80.ffe100.0000.f21a.3596.0020481a3596.f0 -peerlist
“47.0005.80.ffe100.0000.f21a.10bb.0020481a10bb.90
47.0005.80.ffe100.0000.f21a.3552.0020481a3552.10
47.0005.80.ffe100.0000.f21a.3218.0020481a3218.44”
The -peerlist is a space-separated list of ATM

NOTE addresses of DLE peers. Therefore, enclose the
list with either quotation marks “ ” or with curly
braces { }.

This command creates a co-located LES and BUS

NOTE using a single AMI command. You cannot create
a BUS separately using ForeThought 5.0 or
greater.
If no -type is entered, the switch assumes

NOTE Ethernet and uses 1516 as the -mtu size. If
token-ring is used as the type, but no -mtu
size is entered, 4544 is used as the size.
You must enter the address of each of the DLE

NOTE peer servers when you are starting DLE; e.g., if
you want four peers, then all four must be
configured with the addresses of the other three
peers, as well as their own LES address, at the
time that each LES/BUS is started. If you want to
add new peers to the list at a later time, you can
use the services lane les peers new
command.
2. Use the following AMI command to verify that the LES and the BUS have been
started and are running.
services lane les show
Admin Oper ELAN

Index State State Type MTU Name SECURE FwdTLVs FwdARPs ID
1 up up ethernet 1516 engineering disabled enabled disabled 100
TRNUMBER: N/A
LES: 0x47.0005.80.ffe100.0000.f21c.10bb.0020481c10bb.10
BUS: 0x47.0005.80.ffe100.0000.f21c.10bb.0020481c10bb.10
LECS: 0xc5.0079.00.000000.0000.0000.0000.00a03e000001.00
ANYCAST: c5000580ffe1000000f21c10bb0020323c10bb10
PEER: 0x47.0005.80.ffe100.0000.f21c.10bb.0020481c10bb.10
PEER: 0x47.0005.80.ffe100.0000.f21c.10bb.0020481a42bb.10
PEER: 0x47.0005.80.ffe100.0000.f21c.10bb.0020482a55a0.70
The OperStatus field shows up, meaning that the LES and the BUS are running.

3. Configure the DLE peer servers by opening an AMI session on one of the other
switches that will run a DLE peer server and enter something similar to the
following:
services lane les new -atmaddress 10 -name engineering -anycast

“47.0005.80.ffe100.0000.f21a.3552.0020481a3552.10
47.0005.80.ffe100.0000.f21a.10bb.0020481a10bb.90
47.0005.80.ffe100.0000.f21a.3218.0020481a3218.44”
4. Open an AMI session on the third switch that will run a DLE peer server and enter
something similar to the following:
services les new -atmaddress 44 -name engineering -anycast

“47.0005.80.ffe100.0000.f21a.3218.0020481a3218.44
47.0005.80.ffe100.0000.f21a.10bb.0020481a10bb.90
47.0005.80.ffe100.0000.f21a.3552.0020481a3552.10”
Once all of the peers have been created, use the services lane les advanced-
info command to verify that the peers have established point-to-point and point-
to-multipoint connections to each other.
Configuring an
Emulated LAN

3.6.4 Starting the LEC(s) and Joining an ELAN

Now that the ELAN services have been started, you can have LECs join the ELAN that you
have created.
The switch software only allows you to create an

NOTE instance of a LEC on a switch. To create an
instance of a LEC on a host, you must use the
ForeRunner VLAN Manager or use a ForeRunner
host adapter. Please refer to the respective User’s
Manual for instructions.
A LEC created on the switch cannot join a Token

NOTE Ring ELAN. It can only join an Ethernet ELAN.
1. To start a LEC that will attempt to join the ELAN, use the following AMI command
on the switch that is going to be a LEC:
interfaces lec new
Usage:
[-atmaddress] <NSAP Selector> LEC ATM Address
[[-adminstatus] (up|down)] Admin State
[[-mode] (wellknown|manual)] LEC Config Mode (default: wellknown)
[[-lecs] <NSAP Address>] LECS ATM Address
[[-les] <NSAP Address>] LES ATM Address
[[-ipaddr] <IP Address>] Interface IP Address
[[-netmask] <IP Address>] Interface Netmask
[[-ifstate] (up|down)] Interface State
The recommended (and default) method for

NOTE starting a LEC is to use the wellknown mode,
meaning that the LEC will attempt to contact the
LECS on the “well-known” address as defined
by the ATM Forum’s LAN Emulation standards
(47.0079.00.000000.0000.0000.0000.00
A03E000001.00).

For example, to start a LEC that attempts to join the ELAN called engineering,
enter the following:
interfaces lec new -atmaddress 2 -name engineering -ipaddr

192.168.61.25 -netmask 255.255.255.0
If you want to use the manual mode, you must

NOTE enter either a LECS address other than the well-
known address or you must enter a LES address.
If you enter a LES address, this means that the
LEC bypasses the LECS and directly contacts the
specified LES.
In order to use DLE, the LES address to be used

NOTE must be the anycast address of the DLE peer
server, not the server NSAP address.
2. Verify that the LEC has joined the ELAN by using the following AMI command:
interfaces lec show
Admin Oper
Index State State Mode MACaddress IfName IfState ELAN
1 up up wellknown 00:20:48:15:09:6b el0 up engineering
IpAddress: 198.168.61.25
Netmask: 255.255.255.0
LEC: 0x47.0005.80.ffe100.0000.f21c.09f1.0020481c09f1.00
Configuring an
Emulated LAN
LECS: 0x47.0079.00.000000.0000.0000.0000.00a03e000001.00
LES: c000580ffe1000000f21c09f10020481c09f1f0
The OperStatus field shows up, meaning that the LEC has successfully joined
the ELAN.
If the OperStatus field shows joining, this

NOTE means that the LEC is still registering with the
ELAN. Wait a few seconds and check it again.
When it has finished, the OperStatus field
displays up.

You can also use the command interface lec advanced-info to display more
information about the LEC.
3. After the first LEC has joined the ELAN, you can perform the same steps in this
section on another switch to allow a LEC to run on that switch. You can also use
the VLAN Manager or the host software to add more LECs to this ELAN. Once all
the LECs have joined, the ELAN is complete.

3.7 Upgrading an ELAN to Use DLE

This section describes how to upgrade your ELAN if you had previously configured LANE,
you want to upgrade some or all of the clients from ForeThought 4.1.x to ForeThought 6.x, and
you want to upgrade all the equipment that is running services to ForeThought 6.x using DLE.
If you used the VLAN Manager to create your

NOTE ELAN, you should use it to upgrade the ELAN.
Refer to the ForeView VLAN Manager User’s
Manual for the appropriate steps.
The following basic steps are involved in upgrading your ELAN to use DLE. Each of these
steps is described in detail in the following sections. It is recommended that you read the
entire section before attempting to upgrade to DLE. These steps describe the commands for a
switch. Refer to the corresponding User’s Manual for the appropriate steps for commands for
hosts, PowerHubs, ES-3810s, or the VLAN Manager.
1. Edit the LECS.CFG file.
2. Delete the old LES and BUS.
3. Upgrade the switches that are running services.
4. Create the DLE peer servers.
5. Transfer the updated LECS.CFG file.
6. Restart the LECS.
7. Recreate the LECs.
8. Create the last DLE peer.
9. Add the last DLE peer to each list of peers.
10. Change the last failover ELAN information in the LECS.CFG file.
Configuring an
Emulated LAN
11. Transfer the final LECS.CFG file.
12. Restart the LECS.

3.7.1 Edit the LECS.CFG File

Before you edit the LECS.CFG file, you may wish
NOTE to back it up to a host using the system
filesystem put -filename <text> -url
<URL> command. If you are using TFTP as the
transfer protocol (this is the default), the remote
host to which the FLASH file will be sent must
be running the TFTP server code and must have
on an empty file in the /tftpboot directory on
the remote host to receive the FLASH file. See
Chapter 4 in Part 3 of the AMI Configuration
Commands Reference Manual for more
information. If you are using FTP as the transfer
protocol, you only need to enter the command
shown below to perform the transfer.
Transfer the LECS.CFG file that is currently being used by the LECS to a workstation (other
than the one to which you backed up the file) so you can edit the file.
myswitch:-> system filesystem put -filename lecs.cfg -url 169.144.85.195/

tftpboot/lecs.cfg
Transferred 2323 bytes of fs:/lecs.cfg
The information about the LES addresses in your old LECS.CFG file may look something like
this before editing:
Match.Ordering: Mktg|0, Mktg|1
Mktg|0.Address: 47000580ffe1000000f21a24f90020481a24f900
Mktg|0.Accept: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Mktg|1.Address: 47000580ffe1000000f21a390e0020481a390e00
On the workstation that has the LECS.CFG file, use a text editor to make the following
changes:
1. Add a new ELAN to the beginning of the Match.Ordering list that is named like
the old failover ELANs, but without the |n. In this example, there are currently
two failover ELANS named Mktg|0 and Mktg|1. Therefore, add an ELAN named
Mktg.

2. Give the anycast address of the DLE peer servers to the new ELAN. Also, replace
the LES address of the old ELANs with this anycast address, except for the last of
the |n failover ELANs (in this case, Mktg|1). Leave Mktg|1 with old LES address
for now so the LECs can use it until they are all changed over. (It will be changed
at the end of the process.)
3. If there were any non-colocated BUSs (none shown in this example), delete any
lines that refer to them, since each BUS will co-located with a LES after you
upgrade to ForeThought 6.x.
After following the steps above, the information about the LES addresses in your LECS.CFG
file would look like this:
Match.Ordering: Mktg, Mktg|0, Mktg|1
Mktg.Address: c5000580ffe1000000f21c126b0020481c126b66
Mktg.Accept: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Mktg|0.Address: c5000580ffe1000000f21c126b0020481c126b66
3.7.2 Delete the LES and BUS

Display the LES information so that you can find the index number of the LES.
myswitch:services lane les-> show
Configuring an
Emulated LAN
Admin Oper ELAN
1 up up ethernet 1516 Mktg|0 disabled enabled disabled 0
LES: 0x47.0005.80.ffe100.0000.f21a.24f9.0020481a24f9.00
BUS: 0x47.0005.80.ffe100.0000.f21a.24f9.0020481a24f9.00
LECS: N/A
ANYCAST: N/A
PEER: N/A
Administer the LES and BUS down for Mktg|0 using the following AMI command:
myswitch:services lane les-> modify -index 1 -adminstatus down

Delete the co-located LES and BUS.
myswitch:services lane les-> delete -index 1
Verify the LES has been deleted.

No LES information is available.
Do the same for each old LES and BUS, except for the last LES and BUS (in this case, Mktg|1).
All clients will still be using the last LES/BUS temporarily until the DLE peer servers have
been established. The last LES and BUS will be changed over at the very end of the process.
3.7.3 Upgrade the Switches Running Services

Upgrade each of the switches that is going to be running services to ForeThought 6.1.x using
the following AMI command:
oper upgrade <remotehost>:<full path to remotefile>
Do not upgrade the switch that is running the last LES and BUS (in this case, Mktg|1). The
LECs will still be using the last LES/BUS temporarily until the DLE peer servers have been
established.
3.7.4 Create the DLE Peer Servers

You must enter the address of each of the DLE
d
NOTE peer servers when you are starting DLE; e.g., if

you want four peers, then all four must be
configured with the addresses of the other three
peers, as well as their own LES address, at the
time that each LES/BUS is started. If you want to
add new peers to the list, you can use the
services lane les peers new command.
Create a DLE peer server (LES/BUS pair) with the new ELAN name (in this case, Mktg) on
each switch that is to become a DLE peer server.
myswitch:services lane les-> new -atmaddress 0x00 -elanname Mktg -anycast

c5.0005.80.ffe100.0000.f21c.126b.0020481c126b.66 -peerlist
47.0005.80.ffe100.0000.f21a.24f9.0020481a24f9.00
47.0005.80.ffe100.0000.f21c.10bb.0020481c10bb.90

Use the show command to verify the information that you entered:

Admin Oper ELAN
1 up up ethernet 1516 Mktg disabled enabled disabled 100
LES: 0x47.0005.80.ffe100.0000.f21a.24f9.0020481a24f9.00
BUS: 0x47.0005.80.ffe100.0000.f21a.24f9.0020481a24f9.00
LECS: N/A
ANYCAST: c5.0005.80.ffe100.0000.f21c.126b.0020481c126b.66
PEER: 0x47.0005.80.ffe100.0000.f21c.10bb.0020481c10bb.90
Create each of the other peer servers using the same anycast address and giving them the peer
address(es) of each peer in the ELAN. (In this example, there is only one peer.)
myswitch:services lane les-> new -atmaddress 0x90 -elanname Mktg -anycast

c5.0005.80.ffe100.0000.f21c.126b.0020481c126b.66 -peerlist
47.0005.80.ffe100.0000.f21c.10bb.0020481c10bb.90
47.0005.80.ffe100.0000.f21a.24f9.0020481a24f9.00
Use the show command to verify the information that you entered:

Admin Oper ELAN
1 up up ethernet 1516 Mktg disabled enabled disabled 100
LES: 0x47.0005.80.ffe100.0000.f21c.10bb.0020481c10bb.90
Configuring an
Emulated LAN
BUS: 0x47.0005.80.ffe100.0000.f21c.10bb.0020481c10bb.90
LECS: N/A
ANYCAST: c5.0005.80.ffe100.0000.f21c.126b.0020481c126b.66
PEER: 0x47.0005.80.ffe100.0000.f21a.24f9.0020481a24f9.00

3.7.5 Transfer the Updated LECS.CFG File

Transfer the updated LECS.CFG file back to the switch that is running the LECS.
myswitch:system-> filesystem get -url 169.144.85.195/tftpboot/lecs.cfg

-filename lecs.cfg
3.7.6 Restart the LECS

Administer the LECS down and back up again. This forces any active clients to re-establish
their connection with the LECS and forces the LECS to read and use the new LECS.CFG file.
myswitch:services lane lecs-> modify -index 1 -adminstatus down
myswitch:services lane lecs-> modify -index 1 -adminstatus up
Use the show command and look at the OperStatus field to verify that the LECS has come
up again.
myswitch:services lane lecs-> show

LECS: 0x47.0005.80.ffe100.0000.f21a.5208.0020481a5208.f1
WellKnown: 0xc5.0079.00.000000.00000000000000a03e000001.00
Default LES: 0x00000000000000000000000000000000000000f0
3.7.7 Recreate the LECs

If you are going to upgrade any or all of the clients to ForeThought 6.1.x, you should do so now.
Use the following AMI command to upgrade any of the clients that are running on switches:
The rest of these instructions apply whether you

NOTE upgraded the LEC on the switch to ForeThought
6.x or not. For commands to recreate LECs that
are on platforms other than a switch, refer to the
appropriate User’s Manual.

Use the following command to get the interface name from the IfName field as follows:
myswitch:interfaces lec-> show

Admin Oper
1 up up wellknown 00:20:48:1a:2c:78 el17 up Mktg|0
IpAddress: 172.14.15.19
Netmask: 255.255.0.0
LEC: 0x47.0005.80.ffe100.0000.f21c.09f1.0020481c09f1.00
LECS: 0x47.0079.00.000000.0000.0000.0000.00a03e000001.00
LES: 0x47.0005.80.ffe100.0000.f21c.09f1.0020481c09f1.f0
Configure that LEC down.
myswitch:interfaces lec> modify -index el17 -adminstatus down
Delete all instances of the LEC (including anything that had been <ELAN name>|n).
myswitch:interfaces lec-> delete -index 1
myswitch:interfaces lec-> delete -index 2
Recreate the LEC. (You now only need one instance per LEC and you will not specify |n in the
name anymore.)
Configuring an
Emulated LAN
myswitch:interfaces lec-> new -atmaddress 0x11 -name Mktg -ipaddr 192.168.61.25
-netmask 255.255.255.0 -ifstate up

Use the following command to verify that the LEC has joined the ELAN by looking at the
OperStatus field as follows:
myswitch:interfaces lec-> show

Admin Oper
1 up up wellknown 00:20:48:1a:2c:78 el17 up Mktg|0
IpAddress: 172.14.15.19
Netmask: 255.255.0.0
LEC: 0x47.0005.80.ffe100.0000.f21c.09f1.0020481c09f1.00
LECS: 0x47.0079.00.000000.0000.0000.0000.00a03e000001.00
LES: 0x47.0005.80.ffe100.0000.f21c.09f1.0020481c09f1.f0
Repeat all of the steps in this section for each LEC. Refer to the corresponding User’s Manual
for the appropriate steps for re-creating LECs that are running on hosts, PowerHubs, or
ES-3810s. At this time, you may also add any new clients to the ELAN.
3.7.8 Create the Last DLE Peer

After all of the LECs have been changed over and any new LECs have been added, go back
and change the remaining failover LES to a DLE peer. Create the DLE peer as follows:
myswitch:services lane les-> new -index 1 -atmaddress 0x02 -elanname Mktg

-anycast c5.0005.80.ffe100.0000.f21c.126b.0020481c126b.66 -peerlist
47.0005.80.ffe100.0000.f21a.24f9.0020321a26e5.02
47.0005.80.ffe100.0000.f21a.24f9.0020481a24f9.00
47.0005.80.ffe100.0000.f21c.10bb.0020481c10bb.90
3.7.9 Add the Last DLE Peer to Each Peer List

Now add the new peer to the list of each of the existing peers using the following command
on each of those peers:
myswitch:services lane les peers-> new -index 1 -peer

47.0005.80.ffe100.0000.f21a.24f9.0020321a26e5.02

3.7.10 Update the LECS.CFG File

After you add the last DLE peer, go back and delete the old remaining failover ELAN informa-
tion from the LECS.CFG file. Transfer the file to a workstation for editing.
myswitch:-> system filesystem put -url 169.144.85.195/tftpboot/lecs.cfg

On that workstation, use a text editor to delete all lines referring to any of the |n failover
ELANs (in this case, Mktg|0 and Mktg|1). If the LES address portion of your LECS.CFG file
looked like this before editing:
Match.Ordering: Mktg, Mktg|0, Mktg|1
Mktg|0.Address: c5000580ffe1000000f21c126b0020481c126b66
It will look like this after editing:
Match.Ordering: Mktg
Configuring an
Emulated LAN
3.7.11 Transfer the Final LECS.CFG File

Transfer the final LECS.CFG file back to the switch that is running the LECS.
myswitch:system-> filesystem get -url 169.144.85.195/tftpboot/lecs.cfg

-filename lecs.cfg

3.7.12 Restart the LECS

Administer the LECS down and back up again. This forces the active LECs to re-establish their
connection with the LECS and forces the LECS to read and use the new LECS.CFG file.
myswitch:services lane lecs-> modify -index 1 -adminstatus down
myswitch:services lane lecs-> modify -index 1 -adminstatus up
Use the show command and look at the OperStatus field to verify that the LECS has come
up again:
myswitch:services lane lecs-> show

LECS: 0x47.0005.80.ffe100.0000.f21a.5208.0020481a5208.f1
WellKnown: 0xc5.0079.00.000000.00000000000000a03e000001.00
Default LES: 0x00000000000000000000000000000000000000f0
The transfer of your ELAN to DLE services is now complete.

3.8 Upgrading an ELAN without Using DLE

This section describes how to upgrade your ELAN if you had previously configured LANE,
you want to leave the clients running ForeThought 4.1.x, and you want to upgrade all the
equipment that is running services to ForeThought 6.x without using DLE.
The following basic steps are involved in upgrading your ELAN ForeThought 6.x without
using DLE. Each of these steps is described in detail in the following sections. It is recom-
mended that you read the entire section before attempting the upgrade.
1. If the LES and BUS are not co-located, they must be deleted and re-created as
co-located because ForeThought 6.x on the switch only supports co-located services.
If the LES and BUS are already co-located, skip to step 2.
2. Upgrade the switches that are running services.
3. Re-create the LES and BUS if they were not co-located. If the LES and BUS were
already co-located, skip to step 4.
4. Administer up the switches that are running services.
If you used the VLAN Manager to create your

NOTE ELAN, you should use it to upgrade the ELAN.
Refer to the ForeView VLAN Manager User’s
Manual for the appropriate steps.
These steps describe commands for services

NOTE running on a switch. Refer to the corresponding
User’s Manual for the appropriate steps for
Configuring an
Emulated LAN
commands for any services that are running on
hosts, PowerHubs, or ES-3810s.

3.8.1 Deleting the Non Co-located Services

There are several steps involved in changing the non co-located services.
3.8.1.1 Administer Down the Services

Administer down the LECS using a pre-ForeThought 6.x version of software.
myswitch::configuration lane> lecs admin <LECS index> down
Administer down the LES and BUS using a pre-ForeThought 6.x version of software.
myswitch::configuration lane> les admin <LES index> down
myswitch::configuration lane> bus admin <BUS index> down
3.8.1.2 Delete the Non Co-located LES and BUS

Delete the LES and BUS.
myswitch::configuration lane> les delete <LES index>
myswitch::configuration lane> bus delete <BUS index>
3.8.1.2.1 Edit the LECS.CFG File

Before you edit the LECS.CFG file, you may wish
NOTE to back it up to a host using the oper flash put
command. If you are using TFTP as the transfer
protocol (this is the default), the remote host to
which the FLASH file will be sent must be
running the TFTP server code and must have an
empty file in the /tftpboot directory on the
remote host to receive the FLASH file. See
Chapter 4 in Part 3 of the AMI Configuration
Commands Reference Manual for more
information. If you are using FTP as the transfer
protocol, you only need to enter the command
shown below to perform the transfer.

Transfer the LECS.CFG file that is currently being used by the LECS to a workstation (other
than the one to which you backed up the file) so you can edit the file.
myswitch::operation> flash put lecs.cfg 169.144.85.195:/tftpboot/lecs.cfg

On the workstation that has the LECS.CFG file, use a text editor to delete any lines that refer to
any non-colocated BUSs.
After you have made the changes, transfer the LECS.CFG file back to the LECS.
myswitch::operation flash> get 169.144.85.195:/tftpboot/lecs.cfg lecs.cfg
3.8.2 Upgrade the Switches Running Services

Upgrade the switch(es) running each of the services to ForeThought 6.1.x using the following
AMI command:
myswitch::operation> upgrade <remotehost>:<full path to remotefile>
3.8.3 Recreate the LES and BUS Together

If your LES and BUS are already co-located, skip to step 4. If your LES and BUS were not
co-located, re-create each LES and BUS using the following single AMI command:
myswitch:services lane les-> new

Configuring an
Emulated LAN
[-atmaddress] <NSAP Selector> LES ATM address
[[-adminstatus] (up|down)] Admin State (default: up)
[[-bus] <NSAP Address>] BUS ATM Address
[[-type] (ethernet|token-ring)] LAN Type (default: ethernet)
[[-mtu] <mtu>] MTU (default: 1516)
[[-lecs] <NSAP Address>] Secure LECS Address
[[-secure] (enabled|disabled)] Secure
[[-fwdtlvs] (enabled|disabled)] Forward TLVs
[[-elanid] <integer>] ELAN ID
[[-trnumber] <Hex Integer>] Token Ring Number
[[-fwdarp] (enabled|disabled)] Forward ARP
[[-anycast] <NSAP Address>] Anycast ATM Address
[[-peerlist] <text>] Peer ATM Addresses

You need to specify the LES selector byte (the BUS uses the same selector byte by default) and
you need to give the same LES name that you used before. You may optionally specify any of
the parameters, except for the anycast address and the peer addresses since you are not using
DLE.
3.8.4 Administer the Services Up

Administer up the LES/BUS pair and the LECS on the switch(es) running each of the services.
myswitch:services lane les-> -index <integer> -adminstatus up
myswitch:services lane lecs-> -index <integer> -adminstatus up
The upgrade of your ELAN is now complete.

CHAPTER 4 MPOA
This chapter provides an overview of LAN Emulation (LANE) and Multi-Protocol Over ATM
(MPOA) as implemented in FORE Systems’ ForeThought software.
4.1 Overview of LANE/MPOA

FORE System’s ForeThought software is compliant with the ATM Forum’s LAN Emulation Over
ATM Version 1.0 specification. LANE allows higher level protocols and LAN applications to
interoperate, without modifications, with an ATM network.
The LANE components, running on the ATM network, interact to emulate an Ethernet or
Token Ring LAN. This emulated Ethernet or Token Ring LAN is called an emulated LAN
(ELAN). The ELAN components resolve MAC addresses to ATM addresses, replace the con-
nectionless operation of legacy LANs with point-to-point connections, and provide broadcast
and multicast services. The ELAN consists of a LANE/MPOA Client (LEC/MPC) running on
each host in the ELAN, and the following LANE Services:
• the LAN Emulation Server (LES)
• the Broadcast and Unknown Server (BUS)
• the LAN Emulation Configuration Server (LECS)
The LANE services may operate on a FORE Systems ATM switch, PowerHub 7000, PowerHub
8000, ASN-9000, or Solaris workstation. ForeThought also provides support for Distributed
LAN Emulation (DLE) which provides load-sharing and improved fault-tolerance within an
ELAN.
MPOA

MPOA
4.2 LANE Primer

LANE is the foundation on which MPOA is built. Therefore, before presenting an explanation
of MPOA, an understanding of LANE components and their operation in an ELAN is needed.
4.2.1 LANE Components

An ELAN includes the following components:
LANE/MPOA Client (LEC/MPC) The LEC/MPC can wear two different “hats.” When
wearing its LEC “hat,” it simply communicates with
other ELAN components (the LES and BUS) to
resolve MAC addresses into ATM addresses. When it
puts on its MPC “hat,” the additional function of the
LEC/MPC in an MPOA-aware network is to source
and sink internetwork shortcuts.
LAN Emulation Runs on a Solaris workstation or a FORE Systems
Configuration Server (LECS) ATM switch. Maintains information about all ELANs
within the administrative domain. When the
LEC/MPC successfully communicates with the
LECS, the LECS provides a list of ELANs which the
LEC/MPC can join. The LECS may be configured
with various MPOA parameters. LEC/MPCs that
connect to LANE/MPOA services through an
MPOA-aware LECS are configured with these
centrally-supplied MPOA parameters. LEC/MPCs
that connect through an LECS that does not contain
MPOA parameters still perform flow analysis and
attempt inter-ELAN shortcuts according to their
user-editable or factory-default settings.

MPOA
LAN Emulation Server (LES) Runs on a PowerHub 7000, a PowerHub 8000, an

ASN-9000, a FORE ATM switch, or a Solaris
workstation. Maintains information about the
LEC/MPCs within a single ELAN and performs
address resolution. The LES can be configured to
support or disable MPOA operation in an ELAN.
The LES accepts MPOA parameters from registering
LEC/MPCs and MPSs, and also distributes MPOA
parameters to LEC/MPCs in response to queries.
(This is the mechanism used by LEC/MPCs to
determine whether routers in the ELAN are
MPOA-aware).
Broadcast and Unknown Server Runs on a PowerHub 7000, a PowerHub 8000, an
(BUS) ASN-9000, a FORE ATM switch, or a Solaris
workstation. Provides services within a single ELAN
allowing broadcasts, multicasts, and unknown
unicasts. The BUS is MPOA-ignorant.
4.2.2 An Example LANE Configuration

Figure 4.1 shows an example configuration of a single ELAN in a FORE network. The ELAN
includes:
• PC Workstations, each running a LEC/MPC. Each has a ForeRunner ATM adapter,
the ForeRunner driver for the adapter, and one or more ForeRunner ELAN drivers
installed.
• Two ForeRunner ASX-200BX switches running LESs, BUSs, and LECs. Each switch
is also running an LECS. The LES/BUS pairs are configured as peers under Dis-
tributed LAN Emulation. The peer configuration allows the LECs associated with
a particular LES/BUS automatically to reconnect to the remaining functional peer
if their “home” LES/BUS fails.
• A PowerHub 7000 running a LEC/MPC, and providing access to non-ATM net-
works.
MPOA

MPOA
ATM
Each PC
PC2 runs a PC5
PC1 PC3 LEC/MPC PC4 PC6
ASX-200BX Runs a LECS, ASX-200BX Runs a LECS,

SWITCH LES/BUS, and SWITCH LES/BUS, and
LEC LEC
PowerHub 7000 Runs a LEC/MPC
Fast Ethernet FDDI

Ethernet
Figure 4.1 - An Example of an ELAN
4.2.2.1 The Initialization Process

Each LEC/MPC goes through the following process when it starts up:
1. The LEC/MPC obtains its own ATM address via address registration. Optionally,
you can manually specify the ATM address.
2. The LEC/MPC establishes a connection to a LECS using an address obtained via
ILMI, a well-known address, or PVC (0,17). Optionally, you can manually specify
the LECS address.

MPOA
3. The LEC/MPC requests the information needed to join a specified ELAN or the
default ELAN. The LECS has information about available ELANs, what ELANs
each LEC/MPC can join, and which ELAN the LEC/MPC should attempt to join
first.
If a LECS is not available, or if you choose not to use it, you can manually specify
the information required to join a specific ELAN.
4. The LEC/MPC contacts the LES associated with the ELAN it wants to join and reg-
isters its MAC-ATM address pair. It also contacts the BUS associated with the
ELAN. At this point, the LEC/MPC and the LES have the information required to
allow this host to communicate with other hosts on the ELAN as if it were an Ether-
net (or Token Ring) network. Refer to the following section for a description of how
the LEC/MPC connects to other hosts on the ELAN.
4.2.2.2 The Connection Process

To send packets to another host on the ELAN:
1. The LEC/MPC calls the LES to map the MAC destination address into an ATM
address. (The LES maintains a mapping table of the address of all LEC/MPCs on
the ELAN.)
2. If the LES finds an entry in its table for the destination MAC address, it returns the
destination ATM address to the LEC/MPC.
3. The LEC/MPC then opens up a point-to-point ATM connection to the destination
host to send the packet.
4.2.2.3 Multicast and Broadcast Packets

The LEC/MPC sends outgoing multicast and broadcast packets to the BUS which uses a
point-to-multipoint connection to send the packets to multiple ATM addresses in the ELAN.
4.2.2.4 Accessing Fast Ethernet and FDDI Networks

Note that the diagram in Figure 4.1 shows dotted lines from the PowerHub 7000 to the Fast
Ethernet and FDDI networks. This shows how the PowerHub can provide access to non-ATM
networks.
MPOA

MPOA
4.2.2.5 Multiple ELANs

It is possible to set up more than one ELAN in a FORE network. For each new ELAN, you
must configure another LES/BUS instance for that LAN. On the access devices, bridge groups
must be used to associate physical ports with ELANs on the ATM side. An end station in the
ELAN with a ForeRunner adapter can connect to up to 16 ELANs simultaneously.
4.2.2.6 Distributed LAN Emulation

To provide greater resilience, support larger ELANs, and support separated clusters of users
in an ELAN, ForeThought software provides Distributed LAN Emulation (DLE). DLE allows
the LES/BUS functions to be distributed among multiple interconnected LES/BUS instances
called peers. In the example ELAN shown in Figure 4.1, the two LES/BUS pairs running in the
switches function as peers in the same ELAN. The LEC/MPCs are distributed such that they
are not all connected to the same server. With this arrangement, should one of the peer servers
fail, the clients connected to the remaining server continue to maintain connectivity; while the
clients that were connected to the failed server automatically re-establish connectivity to the
ELAN within 60 seconds.
4.2.2.7 Automatic ELAN Selection

To simplify configuration of the ELAN, a host is allowed to join an ELAN without specifying
an ELAN name. If the LECS has been configured to provide the required information, and you
do not manually specify an ELAN name to join when you configure the host’s ELAN driver,
the host initially attempts to join the ELAN specified by the LECS. The host successfully joins
the ELAN if the LECS is available, the proper LES address for the ELAN has been specified in
the LECS, and the LES and BUS are available.
4.2.2.8 Intelligent BUS

This feature reduces broadcast traffic by using the MAC address information in the LES.
When an intelligent BUS receives a unicast frame, the BUS first checks the LES’s mapping
table to see if the MAC address is registered there. If it is, the BUS forwards the frame directly
to the destination, instead of broadcasting.

MPOA
4.3 An Introduction to Multi-Protocol Over ATM

MPOA builds upon the foundation of LANE.
4.3.1 LANE Without MPOA

ATM networks co-exist with and support network applications which may not be ATM-
aware. Consequently, ATM protocols are needed to monitor legacy network protocol (IP, IPX,
Appletalk, etc.) packets and perform translation into ATM cells and circuits. This monitoring
and translation can be performed in one of the following ways:
• in a host protocol stack after packet construction and before packet transmission
• in a LAN-to-ATM edge device as packets move through the network
LANE is one example of such a protocol. It resolves datalink layer addresses into ATM
addresses and establishes circuits to the destination addresses. Network addresses within a
subnet can be learned using LANE’s broadcast support.
However, LANE relies on physical routers to deliver packets across subnets (see Figure 4.2).
Because routers must examine – and modify – every packet, ATM cells must be reassembled
into packets, modified, and re-segmented at every router hop. This process imposes significant
transmission delays between the source and destination of the network traffic.
Router
ELAN
ELAN marketing
engineering
ATM cloud
MPOA
Figure 4.2 - LANE without MPOA

MPOA
In addition to LANE, protocols such as IP can operate over an ATM network via the IETF
Internetworking Over NBMA Networks (ION) Working Group’s Next Hop Resolution Protocol
(NHRP). NHRP allows the ATM network to be divided into Logical IP Subnets (LISs). Using
NHRP, routers are still required to interconnect these subnets; but NHRP permits intermediate
routers to be bypassed on the data path. NHRP allows entities called Next Hop Clients (NHCs)
to send queries between different subnets. These queries are propagated using Next Hop Serv-
ers (NHSs) via paths found using standard routing protocols. Consequently, NHRP enables
the establishment of VCC data paths across subnet boundaries without requiring physical routers
in the data path.
4.3.2 Why MPOA?

The ATM Forum developed the Multi-Protocol over ATM (MPOA) specification to address the
limitations of LAN Emulation. MPOA extends ATM support of legacy networks into the net-
work layer. The main objective of MPOA is the efficient transfer of unicast data between sub-
net(s).
MPOA introduces LANE/MPOA Clients (LEC/MPCs) and MPOA Servers (MPSs) and
defines the protocols that are required for LEC/MPCs and MPSs to communicate. LEC/MPCs
issue queries for ATM addresses, and receive replies from the MPS using these protocols.
MPOA also maintains interoperability with the existing infrastructure of routers. MPOA Serv-
ers reside in routers that run standard Internetwork Layer routing protocols such as OSPF,
thus providing integration with existing networks.
ForeThought software implements MPOA shortcuts for IP traffic. It does this by adding capabil-
ities to LANE, not by replacing LANE. LANE/MPOA client drivers are extended LANE driv-
ers. When handling traffic within the same ELAN and subnet, they function like LECs.
However, when handling traffic that crosses subnets, LEC/MPCs initially work with MPOA
servers (MPSs) to use MPS-established hop-by-hop circuits. Then, for traffic flows that exceed
configurable limits, shortcut circuits are built that allow the traffic to traverse the route without
the necessity of the router(s):
• reassemble packets from ATM cells
• modify the packets
• and then re-segment the packets for transmission to the next hop
Consequently, traffic flowing through a shortcut VCC moves at essentially wire speed from
source to destination (see Figure 4.3).

MPOA
ELAN ELAN
engineering marketing
Router
Shortc
ut VCC
ATM cloud
Figure 4.3 - LANE with MPOA
MPOA

MPOA
4.3.3 MPOA Components

MPOA requires LANE services for both ELAN traffic handling and MPOA configuration.
LANE/MPOA Client (LEC/MPC) The LEC/MPC can wear two different “hats.” When
wearing its LEC “hat,” it simply communicates with
other ELAN components (the LES and BUS) to
resolve MAC addresses into ATM addresses. When it
puts on its MPC “hat,” the additional function of the
LEC/MPC in an MPOA-aware network is to source
and sink internetwork shortcut circuits. A LEC/MPC
that is the source of a shortcut is known as an ingress
LEC/MPC. A LEC/MPC that is the sink of a shortcut
is known as an egress LEC/MPC. The LEC/MPC
includes an NHRP Client (NHC).
An ingress LEC/MPC monitors traffic flow that is

being forwarded over an ELAN to a router that
contains an MPS. When the ingress LEC/MPC
recognizes a flow rate (configurable) that could
benefit from a shortcut (and thus bypass the routed
path), it requests a shortcut to the destination. If a
shortcut is available, the ingress LEC/MPC sets up a
shortcut VCC, and forwards traffic for the
destination over the shortcut.
An egress LEC/MPC receives internetwork traffic

from other LEC/MPCs to be forwarded to its local
interfaces/users. For traffic received over a shortcut,
the egress LEC/MPC adds the appropriate
encapsulation and forwards them via a LAN
interface (that may be a bridge port, an internal host
stack, etc.).
MPOA Server (MPS) An MPS includes an NHRP Server (NHS) and is the
logical component of a router that provides
internetwork layer forwarding information to
LEC/MPCs. The MPS answers MPOA queries from
ingress LEC/MPCs and provides encapsulation
information to egress LEC/MPCs.
The MPS also converts between MPOA requests and

replies, and NHRP requests and replies on behalf of
LEC/MPCs.

MPOA
4.3.4 MPOA Example

The following are the basic requirements for establishing a shortcut across an MPOA-enabled
network:
• There must be LEC/MPCs at each end of the network between which a shortcut is
desired.
• The local router interface at each end must be running an MPS.
• A Next Hop Resolution Protocol (NHRP) path must exist between MPSs.
The following example illustrates a typical ATM network that allows MPOA shortcuts to be
employed.
Each of
PC these runs
a LEC/MPC ELAN
PC PC
engineering
Runs a LECS,
ASX-200BX LES, and BUS
Runs
an MPS
ATM
and a PowerHub 7000
Cloud
LEC/MPC
Runs an MPS and

PowerHub 7000 a LEC/MPC
Runs a LECS,
ASX-200BX LES, and BUS
ELAN
marketing Each of
these runs
PC PC a LEC/MPC
PC
MPOA
Figure 4.4 - MPOA Example Network

MPOA
4.3.4.1 MPS Configuration

The network administrator must configure each MPS with the site-specific IP address match-
ing the gateway address being used by LEC/MPCs in its ELAN.
The MPS on each PowerHub 7000 is configured as follows:
1. For each LANE/MPOA virtual port, specify an ELAN name. The LECS configura-
tion must also be updated to allow the MPS to join these ELANs.
2. For each LANE/MPOA virtual port, specify an IP address.
3. Enable LANE/MPOA support.
4. Enable routing.
5. Save the configuration and reboot the MPS if necessary to make the changes
effective.
Router table information need not be configured. The MPS will instead gather this informa-
tion through routing protocol exchanges with other MPS’s.
4.3.4.2 Initialization
When its host boots, each LEC/MPC automatically goes through the following sequence to
establish a connection to the MPS.
1. The LEC/MPC registers via ILMI with the switch to which it is attached.
2. The LEC/MPC connects to an LECS to which it sends its own ATM address and
the name of the ELAN it wishes to join (the ELAN name is an empty string unless
the LEC/MPC has been site-configured with an ELAN name). The LEC/MPC also
supplies a LANE 1.0 compliant parameter identifying itself as an MPOA-aware
client.
3. Next, the LEC/MPC receives the following from the LECS:
- the name of the ELAN to which it is assigned
- the ATM address of the LES for the ELAN it is joining
- the parameters containing the flow detection and shortcut establishment poli-
cies it is to use
4. The LEC/MPC then connects to its assigned LES, and provides the LES with a
parameter identifying itself as MPOA-aware.
5. Finally, the LEC/MPC connects to the ELAN’s BUS.
Once the LANE/MPOA connections are established, third-party network-layer protocol driv-
ers on the host can establish network-layer connectivity. The methods these upper-layer driv-
ers use to determine host IP addresses, default gateway, and backup gateway addresses vary
depending on the third-party product. For example, the LANE/MPOA driver itself permits
these drivers to use BOOTP or DHCP to obtain IP configuration information.

MPOA
4.3.4.3 Flow Analysis

On a LEC/MPC’s host or edge device, IP packets with destinations within the host’s subnet
are sent using LANE 1.0 methods; i.e., the client puts on its LEC “hat” and works directly with
its ELAN’s services to connect with local destinations. Packets destined for remote subnets
cause the LEC/MPC to put on its MPC “hat.” This client is then referred to as an “ingress
LEC/MPC.”
Ingress LEC/MPCs associate destination IP addresses with shortcut circuits. Ingress
LEC/MPCs use configurable parameters called flow descriptors to determine whether and when
to trigger creation of shortcut circuits. The ingress LEC/MPC also monitors the most recent
use of a shortcut circuit to determine when to tear down the shortcut. Specifically, when an
ingress LEC/MPC sends a packet:
1. If a shortcut circuit already exists to the IP destination, the LEC/MPC sends the
packet over this circuit.
2. If no shortcut circuit exists, the LEC/MPC determines whether shortcuts to this IP
address are allowed. If shortcuts to the destination IP address are not allowed, the
LEC/MPC sends the packet to the gateway router.
3. If no shortcut circuits exist, and shortcuts to the IP address are allowed, the
LEC/MPC determines if the packet traffic flow exceeds the shortcut enable trigger
value (set by the flow descriptors) for the destination IP address’s flow. If the flow
exceeds the trigger value, the LEC/MPC tries to establish a shortcut circuit to the
destination LEC/MPC (called the egress LEC/MPC). If the flow does not exceed
the trigger value, the ingress LEC/MPC simply sends the packet traffic to the gate-
way router.
4.3.4.4 Making a Shortcut

When the ingress LEC/MPC determines that the packet traffic flow exceeds the shortcut-
enable trigger value, the ingress LEC/MPC tries to establish a shortcut circuit to the egress
LEC/MPC. The following describes how a shortcut is set up:
1. The ingress LEC/MPC initiates the shortcut creation process by sending an MPOA
resolution request to the MPS it uses as a gateway router (this MPS is called the
ingress MPS). The MPS converts the request to a next hop resolution protocol
(NHRP) request. This NHRP request includes the destination’s IP address and asks
for the corresponding ATM destination address.
2. This request is passed along hop-by-hop until it reaches the final MPS (called the
egress MPS) on the route to the destination IP address.
3. The egress MPS sends a cache imposition request to the egress LEC/MPC. The
MPOA
egress LEC/MPC sends a cache imposition reply, which is converted to a NHRP

response by the egress MPS. The ingress MPS converts the NHRP response to an
MPOA resolution response and transmits it to the ingress LEC/MPC.

MPOA
4. When the ingress LEC/MPC receives the NHRP response containing the destina-
tion’s ATM address, it first checks if a shortcut circuit to that ATM address already
exists. If a shortcut circuit to that address already exists, it sends the packets via the
existing shortcut circuit. If no shortcut circuit exists it opens a new shortcut circuit
and begins sending packets over it to the destination.
4.3.4.5 Shortcut Teardown

Application programs and networking protocol stacks are MPOA-ignorant and, therefore, do
not tear down shortcut circuits when the shortcut is no longer needed. Therefore, the MPOA
layer itself tears down seldom-used shortcuts to avoid circuit exhaustion in the client and net-
work. When a shortcut is idle for a period exceeding a set limit, the shortcut is torn down.

ForeThought PNNI
PNNI (Private Network Node Interface or Private Network-to-Network Interface) is an ATM
Forum approved protocol which defines interoperability between private ATM switches.
PNNI defines both the routing and signalling standards for inter-switch interoperability.
This chapter provides an overview of FORE Systems’ pre-standard version of PNNI,
ForeThought PNNI (FT-PNNI), and its use in a multiple-switch network. FT-PNNI is a scalable
routing and signalling protocol used in networks containing multiple FORE switches.
FT-PNNI simplifies large network topologies by organizing the nodes (switches) in that net-
work into smaller groups.
It is this reorganization of the network topology that makes FT-PNNI’s simplified routing pos-
sible. By segmenting a large network into smaller peer groups of nodes, FT-PNNI reduces the
amount of network topology information that those very nodes must maintain.

ForeThought PNNI
5.1 FT-PNNI Routing

The FT-PNNI routing protocol serves to distribute topology and address reachability informa-
tion between switches and groups of switches in a network. This topology and addressing
information is used by switches to compute paths through the network. The functions of the
FT-PNNI routing protocol include the following:
• Hello Protocol
• Topology Database Exchange
• Flooding
• Hierarchical Routing
5.1.1 Hello Protocol

FT-PNNI hello packets are exchanged between neighboring nodes. Each switch (node) trans-
mits a hello indication on each of its FT-PNNI routing channels at regular intervals. The time
between these hello indications is called the Hello Indication Interval. When a switch receives
a hello indication from one of its neighbors, it stores the logical link (loglink) from that neigh-
bor to itself in the topology database.
Based on the hello indications, loglinks are refreshed periodically. Since loglinks discovered as
a result of hello indications are unidirectional, each switch stores unidirectional loglinks with
its immediate neighbor as the source and itself as the destination.
5.1.2 Topology Database Exchange

Each switch sends to each of its neighbors a group of loglinks from its topology database at
regular intervals, called the NSAP map indication interval. This exchange of information
between neighboring switches ensures that the topology databases of the switches stays syn-
chronized.
5.1.3 Flooding
Flooding is a reliable hop-by-hop propagation of loglinks throughout a peer group. Whenever
a new loglink is discovered by a switch, the switch immediately broadcasts the existence of
this link to all of its neighbors. The neighboring switches then broadcast the existence of the
same link to all of their neighbors. Very quickly, the existence of the new loglink is flooded
throughout all of the switches in the network.
Similarly, when a link goes down, or when a significant change is seen in the metrics of a
loglink, the latest state of the loglink is propagated immediately throughout the network.

ForeThought PNNI
5.1.4 Hierarchical Routing
ForeThought PNNI
FT-PNNI operates in a hierarchical topology. The structure of the hierarchy is defined by the
peer group ID in a routing domain. Address assignment in FT-PNNI, therefore, corresponds to
this hierarchical topology, providing increased scalability.
104 bits
Peer Group ID
Switch Summary Prefix
Switch Prefix
Figure 5.1 - Example of a 13-byte Switch Prefix
5.1.4.1 Hierarchical Addressing

FT-PNNI uses private ATM address prefixes (NSAP prefixes) as node identifiers. FT-PNNI
does not distinguish between node identifiers and reachability information. Thus, the IDs of
nodes in the FT-PNNI addressing map are NSAP prefixes. In the default case at the lowest
level peer group, the switches have a 13-byte prefix as their node ID and end systems (hosts)
have a 19-byte prefix as their node ID.
5.1.4.1.1 Switch Prefix

Each switch in a FT-PNNI network is configured with a 13-byte prefix called the switch prefix.
Hosts that are attached to the switch are presented with this prefix during ILMI address regis-
tration. In this way, end systems are configured with a private ATM address that includes the
13-byte switch prefix.

ForeThought PNNI
5.1.4.1.2 Switch Summary Prefix

Each switch is configured with a switch mask (swmask) which gives the length of the switch
summary prefix within the switch prefix. The swmask gives the number of most significant bits
of the switch prefix that constitute the switch summary prefix. Since all end system addresses
attached to a switch have the same switch summary prefix, their reachability information can
be summarized by this prefix (i.e., by the switch summary prefix).
5.1.4.1.3 Peer Group ID

Each switch is configured with a peer group mask (pgmask) which gives the length of the peer
group ID within the switch summary prefix. The pgmask gives the number of most significant
bits of the switch summary prefix that constitute the peer group ID.
Each peer group has a peer group ID that uniquely identifies it from every other peer group.
Every node (switch or end system) in a particular peer group shares that same unique peer
group ID, thereby indicating membership to that peer group.
A simple example of summarizing by peer group ID can be seen in Figure 5.3, where every
switch (and end system, although not shown) in peer group A is identified starting with “A.”
5.1.4.2 Path Computation

Path computation is performed on demand whenever FT-PNNI signalling requests a path to a
given destination. The Bellman-Ford Shortest Path algorithm is used to compute the shortest
path tree of all nodes in the topology with the local node as the source. The administrative
weight metric in the loglinks is used as the minimizing criterion in computing the shortest
path route. In the case of finding multiple equal-cost paths to a given destination, available
cell rate is used to break ties.

ForeThought PNNI
5.2 The Physical Network
ForeThought PNNI
In an ATM network, data is sent and received over virtual circuits, or circuits that only exist
when needed. This communication over these virtual circuits is made possible by signalling
that occurs between the switches in the network.
In a network of FORE switches, any new addition to the topology is recognized immediately
by all nodes (switches) having a direct connection to the new node. Then each switch that has
recognized the new switch sends a message to each of its direct connections, and so on. Even-
tually, and within a very brief period of time, every switch in the network is aware of the new
addition and the links by which that new addition can be reached. This topology is stored by
each switch in its local topology database.
In a small, local area network (LAN), the topology is relatively simple, meaning that the
switches in the LAN have a relatively small topology database to maintain. As the LAN
grows, however, and more switches are added, that same database can grow to be very large
in a short period of time.
As this topology database grows, the amount of information a switch must look up when
searching for an address also grows. In the end, this can result in delayed connection set-up,
congestion in the network, and even lost data.
Figure 5.2 depicts a typical ATM network, containing 21 FORE switches ( ). The hierarchy
of this network is flat, meaning that each switch must be aware of all the other switches in the
network, as well as all the possible routes to those switches. As more switches are added to
this network, the hierarchy will become more complex and the switches will have to contend
with larger topology databases.

ForeThought PNNI
Figure 5.2 - Private ATM Network with 21 Switches and 34 Bidirectional Links
It is in these large, single-level networks that FT-PNNI is most useful, because it lets you sim-
plify large network topologies by creating a two-level hierarchy. In this hierarchy, clusters of
contiguous switches are grouped together and they are collectively summarized by a single,
logical node.
Figure 5.3 shows the same network as in Figure 5.2 after being organized into peer groups,
now having a two-level hierarchy. The subsections that follow explain the organization of
these peer groups, how they simplify the overall network topology, and how they change the
logical view of the network.

ForeThought PNNI
ForeThought PNNI
Peer Group C
Peer Group
C.6
Border Switch Peer Group D
Switch/Node D.4 C.5
Logical Link D.3 C.4
D.2 C.3
(loglink)
D.1 C.2
C.1
A.1 B.2
B.3 B.8
A.2
B.4
A.3
B.1
Peer Group A B.5 B.7
B.6
Peer Group B
Figure 5.3 - Example of FT-PNNI Hierarchy Showing Lowest-Level Peer Groups

ForeThought PNNI
5.2.1 Peer Groups

The FT-PNNI hierarchy begins with a network of switches, organized into peer groups. A peer
group is a collection of interconnected switches that are organized into a group. Peer group
organization can be determined by a network administrator, but switches that are located
close to one another are usually made into a peer group.
The network shown in Figure 5.3 is organized into four lowest-level peer groups: A, B, C, and
D. The switches within a certain peer group are numbered according to that particular group.
For example, the switches in peer group A are identified as A.1, A.2, and A.3.
Peer groups have a peer group identifier (ID), assigned at configuration time and exchanged
in hello indication messages. Neighboring switches can determine if they belong to the same
peer group by comparing these peer group IDs.
Switches in a peer group are aware of the topology of their own peer group and the existence
of all other peer groups. They recognize the links between their peer group and others, but
they are not aware of internal topology information of other peer groups. Instead, the topolo-
gies of other peer groups are summarized as a single, reachable location, known as a peer
group summary node (PGSN).
5.2.2 Peer Group Topology

To maintain an accurate and updated view of its relative location and status, a switch periodi-
cally sends a hello indication message to every other switch with which it has a direct connec-
tion. These hello indications contain the switch prefix, peer group membership information,
and link metrics (attributes) for the physical link between the two switches.
Through this regular exchange of messages, each switch learns which switches are its immedi-
ate neighbors, to what peer groups they belong, and whether or not the link between itself and
its neighbors is valid.
5.2.3 Border Switches

A border switch is any switch that has at least one link to a switch in another peer group. Bor-
der switches play an important role in FT-PNNI because they are responsible for summarizing
reachability information for their respective peer groups, appropriately filtering the flow of
topology database information across peer group boundaries, and building the lowest level
source route for call setups entering the peer group.

ForeThought PNNI
5.2.4 Peer Group Summary Node (PGSN)
ForeThought PNNI
A PGSN is a virtual (logical) or imaginary node that summarizes a peer group’s reachability
information. The PGSN has the peer group ID of its peer group as its switch summary prefix.
Each border switch in the peer group advertises a logical link (loglink) to the PGSN. The
PGSN is a logical representation of the switches contained in a peer group.
5.2.5 Backbone Topology

Loglinks between border switches and loglinks from border switches to PGSNs are called
backbone links and considered part of the backbone topology. Information regarding these
backbone links is propagated across peer group boundaries during database exchange and
flooding.
5.2.6 Single Switch Perspective

The main reason for grouping switches in large networks is to simplify each individual
switch’s view of the topology. For example, each switch in peer group A is aware of every
other switch in peer group A, the border switches in the rest of the network, the links between
them, and the backbone topology. Switches in peer group A are not aware, however, of the
internal topology of other peer groups. Instead, individual switches see a PGSN (see Figure
5.4).

ForeThought PNNI
Backbone Link
Peer Group Summary D C
Node (PGSN)
D.2
C.2
C.1
A.1
B.8
A B.1
A.3 B
A.2
Figure 5.4 - View of the Network from Switches in Peer Group A

In April, 1996, the ATM Forum (ATMF) approved an industry specification for PNNI Phase 1
(Private Network Node Interface or Private Network-to-Network Interface). PNNI (Private
Network Node Interface or Private Network-to-Network Interface) is a scalable protocol that
defines both the routing and signalling standards for interoperability between private ATM
switches. PNNI simplifies large network topologies by organizing the nodes in a network into
ATM Forum PNNI

smaller groups. Through this re-organization, PNNI reduces the amount of topology informa-
tion that those nodes must maintain.
The ATMF PNNI 1.0 Specification requires every PNNI implementation to support peer
group construction and communication within a peer group; support of communication
among peer groups using border nodes, split switches, or PGL hierarchy is optional.
ForeThought 5.x supports the mandatory requirements defined in the ATMF PNNI 1.0 Specifi-
cation, includes some of the defined optional requirements, and includes FORE extensions to
provide improved calls-per-second performance, easier configuration, configurable address
filtering, and scalability to larger networks. ForeThought 5.x also allows you to connect peer
groups to each other using the split-switch feature, but it does not support the optional ATMF
border node or Peer Group Leader (PGL) hierarchy features.
ForeThought 6.x continues the support of the features mentioned above and adds support for
the optional ATMF border node subset, the PGL/Logical Group Node (LGN) subset, and the
border node with LGN subset.
This chapter provides an overview of FORE Systems’ implementation of ATMF PNNI (hereaf-
ter referred to as PNNI), and its use in a multiple-switch network. The following topics are
discussed in this chapter:
• Section 6.1 - PNNI Routing Protocol
• Section 6.2 - PNNI Signalling Protocol
• Section 6.3 - Border Node, PGL, and LGN Support
• Section 6.4 - Internetworking between PNNI and FT-PNNI
• Section 6.5 - Path Extension QoS
• Section 6.6 - Load-balanced Routing
• Section 6.8 - Configuring PNNI

ATM Forum PNNI
6.1 PNNI Routing Protocol

The PNNI routing protocol enables the distribution of topology and address reachability
information between switches and groups of switches. Reachability information is the data that
explains how to contact a single ATM address or a group of ATM addresses that are summa-
rized by a single prefix. This information is used to compute routes across the network. The
key elements of the PNNI routing protocol are as follows:
• Section 6.1.1 - Hello Protocol
• Section 6.1.2 - Database Exchange Protocol
• Section 6.1.3 - Flooding Protocol
• Section 6.1.4 - Path Computation
• Section 6.1.5 - Hierarchical Routing
6.1.1 Hello Protocol

The Hello protocol is used to discover and verify the identity of neighbor nodes and to deter-
mine the status of the logical links to those nodes. Logical links between nodes in the same
peer group are called horizontal links. Hello packets are exchanged at regular intervals over
links connecting neighboring nodes. Initially, these packets are used to exchange identities
between the nodes and to determine if they belong to the same peer group. After that, Hello
packets are used as keep-alive messages for the links over which the packets are exchanged.
After a configured number of hello intervals have passed without receiving a hello message,
the link is declared to be down.
6.1.2 Database Exchange Protocol

When neighboring nodes determine that they are peers; i.e., that they belong to the same peer
group and they are running compatible versions of the PNNI protocol, they begin a database
exchange process. At regular intervals, each node tells the other what information it has in its
topology database and requests similar information from the other neighboring nodes. This
exchange of information ensures that the topology databases of all switches in the network
stay synchronized.

ATM Forum PNNI
6.1.3 Flooding Protocol

Topology information is organized into smallest units called PNNI Topology State Elements
(PTSEs). PNNI routing uses a flooding mechanism to propagate these PTSEs throughout the
network in a hop-by-hop fashion. PTSEs are encapsulated into PNNI Topology State Packets
(PTSPs) for transmission. PTSPs always contain the node identifier (ID) and the peer group ID
of the switch that originated all of the PTSEs.
When a PTSP is received, its component PTSEs are examined. Each PTSE is acknowledged by
sending a PTSE acknowledgment packet (containing possibly a bundle of PTSE acknowledg-
ments) to the sending peer. If the PTSE is more recent than the copy that the receiving node
has in its topology database, the database copy is updated and is re-flooded to all neighboring
ATM Forum PNNI

peers except the one from which the PTSE was received. A PTSE sent to a neighboring peer is
periodically retransmitted until it is acknowledged.
In this fashion, when a link is created or goes down, this information is propagated immedi-
ately throughout the network. Similarly, when a significant change is seen in the metrics of a
link, this state change is flooded throughout the network.
6.1.4 Path Computation

Path computation is done on the source (or originating) switch. All of the network topology
information obtained and updated by the higher layers of the protocol is used to determine a
path through the network which is efficient and which satisfies all of the relevant Quality of
Service (QoS) parameters.
The path is fully specified by the source switch all the way to the destination. Path information
is stored in a DTL, which is forwarded along with the setup message to establish the circuit for
a particular call.

ATM Forum PNNI
6.1.5 Hierarchical Routing

Hierarchical routing allows contiguous switches to be organized into peer groups. The
switches within each peer group exchange detailed topology information about their own
group that is not visible to switches outside the peer group. Similarly, switches within each
peer group do not receive detailed topology information about switches outside of their peer
group. Instead, each peer group is represented externally by its Peer Group Leader (PGL). The
PGLs exchange among themselves summary topology and reachability information pertain-
ing to their respective peer groups.
This idea is illustrated in Figure 6.1. As you can see, the switches in a given peer group know
about the topology information of their own peer group. However, they are not aware of the
individual members of the other peer group. They only know how to reach the other group
via their PGL.
* +
*
*
*View of the network
from a
* switch
*
* +
* +
+
*
+
Figure 6.1 - Peer Groups Summarized into a Single Address

ATM Forum PNNI
Hierarchical routing reduces the number of links that each switch must maintain. For exam-
ple, if the switches in Figure 6.1 were in a flat network, each switch would have to be aware of
all 16 links shown. If the switches are separated into peer groups, the routing table for each
switch becomes smaller.
In the example shown in the inset in Figure 6.1, a * switch only needs to know about the four
links inside its peer group plus one link to the O peer group and one to the + peer group. This
is a total of six links, as opposed to 16. The example in the inset is hierarchical because the *
switches view all of the O switches as a single node and all of the + switches as another single
node.
ForeThought 6.x supports the dynamic configuration of ATMF PNNI PGL hierarchical net-
ATM Forum PNNI

works. (Although ForeThought 5.x does not support the dynamic configuration of ATMF PNNI
PGL hierarchical networks, FORE switches running ForeThought 5.x can be placed in and inter-
act appropriately within PNNI hierarchical PGL networks.) For more information about
ATMF border node capabilities, ATMF PGL/LGN capabilities, and ATMF border node with
LGN capabilities, see Section 6.3. For more information about how ForeThought can be config-
ured to connect multiple ATMF PNNI peer groups, see Section 6.4.1.
6.1.5.1 Peer Groups

A peer group is a group of physically connected nodes that have a portion of their NSAP
address in common. As shown in the example in Table 6.1, the switches share the first portion
of their NSAP prefix.
Table 6.1 - Example of Peer Group Addressing
Peer Group A NSAP Prefix

switch_1 47.0005.80.ffe100.1111.1111.21b8
switch_2 47.0005.80.ffe100.1111.1111.82c3
switch_3 47.0005.80.ffe100.1111.1111.1559
switch_4 47.0005.80.ffe100.1111.1111.aa17
All of the NSAP addresses of the group members can be represented as a single, shortened
(summarized) address. Additionally, every member of a peer group has the same level. Mem-
bers of a peer group are connected to each other by horizontal links.
Each peer group is summarized by its PGL as a single group node. A node is a logical entity
that resides in a switch and performs routing operations such as discovering other nodes in
the network, maintaining a topology database of its peer group, exchanging that database
with its neighbors, and computing paths from itself to other nodes in the network.

ATM Forum PNNI
6.1.5.2 PGL/LGN Hierarchy

As shown in Figure 6.2, each peer group elects one PGL. (See Section 6.3.8.2 for more informa-
tion about PGL elections.) Upon being elected, the PGL activates a LGN within the same
switch. The LGNs represent the peer group at the next higher (parent) level.
The interconnection of LGNs form a higher-level topology. LGNs can, in turn, be grouped into
higher-level peer groups that, in turn, are represented by LGNs in an even higher level topol-
ogy, and so on. LGNs communicate with each other over dynamically established SVCC
RCCs. The LGNs summarize information about the child peer groups they represent and
flood topology information learned at the higher level to the PGL. The PGL, in turn, floods
that information to its peer group.
For example, as shown in Figure 6.2, switch B.1.1 belongs to the lowest-level peer group B.1,
which, in turn, belongs to a higher-level peer group B. Switch B.1.1 has detailed topology
information of peer group B.1, as well as higher-level topology information of peer group B.
Peer groups B.2 and B.3 also belong to peer group B. However, switch B.1.1 knows about the
LGNs representing B.2 and B.3, respectively, although it does not know the detailed topology
inside peer group B.2 or peer group B.3.
PNNI peer group boundary

PNNI node
PNNI link B.2
PG B
PGL
Level 72
LGN
B.1 B.3
border node
B.3.4
B.1.2 B.2.3
B.3.1 B.3.3
B.1.1
Level 80 B.2.1 B.2.2 Level 80
B.1.3 B.3.2 PG B.3
PG B.1 Level 80
PG B.2
outside link
horizontal link (inside link)
Figure 6.2 - PGL Hierarchy Example

ATM Forum PNNI
6.1.5.3 Border Nodes

Border nodes, such as switch B.1.2 in Figure 6.2, have at least one link that crosses a peer
group boundary (an outside link). No flooding takes place on the outside links. The border
nodes only use the outside links for hello messages. The hello messages advertise reachability
across the border. In Figure 6.2, switch B.1.2 advertises that it can reach B.2, while switch B.2.3
advertises that it can reach B.1.
6.1.5.4 Lop-sided Hierarchy

A lop-sided hierarchical network is one in which a lowest level node discovers that it is in a
peer group with one or more of its neighbors’ ancestors; i.e., the lowest level node is acting as
ATM Forum PNNI

a peer to one or more LGNs. As shown in lop-sided hierarchy example in Figure 6.3, switches
B.4, B.5, and B.6 are all lowest level nodes. However, they are part of peer group B which con-
tains parent node LGNs B.1, B.2, and B.3.

PNNI node B.4
B.5
PNNI link Level 72 B.2
PGL PG B
LGN B.6
B.1 B.3
B.3.4
B.1.2 B.2.3
B.3.1 B.3.3
B.1.1
B.2.1 B.2.2 Level 80
Level 80
B.1.3 B.3.2 PG B.3
PG B.1 Level 80
PG B.2
Figure 6.3 - Example of a Lop-sided Hierarchy

ATM Forum PNNI
6.2 PNNI Signalling Protocol

PNNI signalling is used to establish point-to-point and point-to-multipoint connections across
the network. This protocol is based on ATMF UNI signalling with mechanisms added to sup-
port source routing and crankback.
6.2.1 Source Routing and DTLs

A directed Designated Transit List (DTL) is a source route which specifies the preferred call
routing path that the PNNI router should use when setting up an SVC or an SVP. Each DTL is
a source route and each entry in the DTL represents a single hop in that source route. Each hop
is represented by a PNNI node and the logical output port at that node.
ATM Forum PNNI DTLs are created using the routing pnni dtl rows new command as
described in Part 2 of the AMI Configuration Commands Reference Manual. A DTL is a source
route (uniquely identified by a pair of indices: node index and DTL index) which specifies the
preferred call routing for the SVC portion of a directed SPVC and the SVP portion of an SPVP.
Each entry (row) in the DTL represents a single hop in the source route.
Once you have created the DTLs, you must create a table (list of DTLs) using the routing
dtltag new command. In this table, each DTL is assigned a weight which indicates the prior-
ity in which the DTLs are tried, with the highest weight being the first one tried. The weights
are arbitrary (e.g., 10, 20, 30, and 40). You can even assign two or more DTLs with the same
weight. The table is also assigned an arbitrary tag number which uniquely identifies the table.
This tag can be applied on a per-SPVC or per-SPVP basis using the -dtltag option under
connections smartchannel pnni pp new (or modify) or under connections
smartpath new (or modify). Multiple SPVCs or SPVPs can use the same DTL table.
When an SPVC or SPVP is created with a DTL tag assigned to it, the DTL in the corresponding
table with the highest weight is tried first. If this DTL is unusable, the DTL with the next high-
est weight is tried. This process is repeated until a DTL in that table is found that provides a
route to the destination which satisfies the QoS requirements.
If none of the user-specified DTLs in the table can be used, and if the -autoDtl option under
connections smartchannel pnni pp new (or modify) is set to enabled, the call is
dynamically routed if an alternative path can be computed. If that option is set to disabled,
then the call setup fails, but the DTLs are retried periodically.

ATM Forum PNNI
When a DTL fails, a back-off mechanism is used to determine when to retry that particular
DTL. The DTL is not attempted for a user-configured interval in order to allow any old infor-
mation to age out and allow the network to stabilize. This interval is set under connections
smartchannel pnni parameters [-backoffinterval] <integer> for SPVCs. It is set
under connections smartpath parameters [-backoffinterval] <integer> for
SPVPs.
However, you can enable or disable this back-off mechanism on a per-SPVC or per-SPVP
basis. Use the -backoffstatus option under connections smartchannel pnni pp new
(or modify) for SPVCs. Use the -backoffstatus option under connections smartpath
new (or modify) for SPVPs. The default setting for both SPVCs and SPVPs is enabled. If you
disable this mechanism, then instead of waiting for the user-configured back-off interval after
ATM Forum PNNI

a failure, the DTLs are retried at every reroute interval. If the -backoffstatus option is then
changed to disabled using the modify command on an existing SPVC or SPVP, the remain-
ing waiting period for any failed DTLs is cleared and the DTLs are attempted again.
6.2.2 Crankback
During PNNI signalling, a call being processed according to a DTL may encounter a blocked
node or link along the designated route. Crankback allows a partial reroute of such a rejected
call so that it does not have to be cleared the whole way back to the source. Additionally, an
indication of the blockage and subsequent reroute information is sent to the originator of the
DTL.

ATM Forum PNNI
6.3 Border Node, PGL, and LGN Support

ForeThought 6.x supports the ATMF border node subset, the PGL/LGN subset, and the border
node with LGN subset of the ATMF PNNI Specification. A border node is a logical node in a
specified peer group which has at least one link which crosses the peer group boundary. The
border node subset provides the following capabilities:
• Section 6.3.1 - Outside Links and Hello Protocol over Outside Links
• Section 6.3.2 - Origination of Uplinks and Uplink PTSEs
• Section 6.3.3 - Entry and Exit Border Node DTL Handling
• Section 6.3.4 - Entry Border Node Crankback Handling
A PGL is a node in a peer group that collects and assimilates data to represent the entire peer
group as a single node in the form of a parent node. An LGN is a virtual representation of a
lower level peer group as a single node in terms of one level of the PNNI routing hierarchy.
The PGL/LGN subset provides the following capabilities:
• Section 6.3.2 - Origination of Uplinks and Uplink PTSEs
• Section 6.3.5 - SVCC RCC Establishment
• Section 6.3.6 - SVCC RCC Hello Protocol
• Section 6.3.7 - LGN Horizontal Link Hello Protocol and PTSEs
• Section 6.3.8 - PGLs
• Section 6.3.9 - Aggregation Tokens and Link Aggregation
• Section 6.3.10 - Nodal Aggregation
• Section 6.3.11 - Internal and Exterior Address Summarization
• Section 6.3.12 - Anycast Address Scoping
Similarly, the border node with LGN support subset provides the following capabilities:
• Section 6.3.5 - SVCC RCC Establishment
• Section 6.3.6 - SVCC RCC Hello Protocol
• Section 6.3.7 - LGN Horizontal Link Hello Protocol and PTSEs
• Section 6.3.9 - Aggregation Tokens and Link Aggregation
Because the three subsets mentioned above overlap in terms of the capabilities that they pro-
vide, each capability is described once in the following subsections.

ATM Forum PNNI
6.3.1 Outside Links and Hello Protocol over Outside Links

In ForeThought 5.x, only four of the possible states of the PNNI Hello Finite State Machine
(FSM) are implemented. These states include: down, attempt, one-way inside and two-way
inside. In ForeThought 6.x, the FSM states that pertain to outside link support are implemented.
These states include: one-way outside, two-way outside, and common outside.
An outside link is a link to a lowest-level outside node. It is not part of the PNNI topology like
a horizontal link or uplink is, so it is not used in path computation. However, outside link sup-
port is essential for the PNNI LGN hierarchy because uplinks and SVCC RCCs are based on
this support. Outside link support means that two PNNI nodes that belong to two different
peer groups can be connected directly with each other by an outside link. When the two PNNI
ATM Forum PNNI

nodes realize this is an outside link, the corresponding Hello FSMs eventually change to the
common outside state, as long as the two nodes share a common peer group ID at a higher
level of the PNNI hierarchy. Otherwise, they change to the two-way outside state.
6.3.2 Origination of Uplinks and Uplink PTSEs

An uplink connects a border node to an upnode. A border node is node in a peer group that has
at least one link which crosses the peer group boundary. An upnode is a node that represents a
border node’s outside neighbor in the common peer group. The upnode must be a neighbor-
ing peer of one of the border node’s ancestors.
Each physical link that is in the common outside Hello state has a logical uplink irrespective of
the aggregation token associated with the physical link. The origination of uplink PTSEs is
fully implemented as defined in Section 5.10.2 of the ATMF PNNI Specification. An uplink
PTSE is generated as soon as the link reaches the common outside Hello state.
6.3.3 Entry and Exit Border Node DTL Handling

ForeThought 6.x implements the handling of DTLs by entry and exit border nodes. An entry bor-
der node is the first node in a peer group which receives a call over an outside link. An exit border
node is the last node in a peer group to see a call. It sends the call out of the peer group over an
outside link. When a signalling call enters a peer group in a hierarchical PNNI network, the
entry border node may need to add several DTLs to the beginning of the list of existing DTLs in
order to process the call. Additionally, an exit border node may need to remove some DTLs
from the list of existing DTLs when it forwards the call to the next peer group.
6.3.4 Entry Border Node Crankback Handling

ForeThought 6.x implements crankback capabilities for entry border nodes. Because an entry
border node can generate DTLs, the entry border node may have to crank back when a link or
port is blocked in a DTL that it originated.

ATM Forum PNNI
6.3.5 SVCC RCC Establishment

When a PGL receives an uplink advertisement of an uplink from one of the peer group’s bor-
der nodes, that advertisement contains information about the neighboring LGN’s (Upnode’s)
node ID, peer group ID, ATM address, etc. The PGL uses this information to establish an
SVCC RCC to the neighboring LGN. The SVCC RCCs are established as defined in Section
5.5.6.1 of the ATMF PNNI Specification.
6.3.6 SVCC RCC Hello Protocol

As soon as an LGN detects an uplink to a neighboring LGN or to a PNNI node (in a lop-sided
hierarchy), it tries to establish an SVCC RCC so that it can communicate with the other LGN
or PNNI node. Once the SVCC RCC is established, the LGN starts the SVCC-based RCC Hello
Protocol as defined in Section 5.6.3.1 of the ATMF PNNI Specification. When the Hello state
reaches the twoWayInside state, it triggers an AddPort event in the neighboring peer state
machine. When this occurs, database summary packets, PTSE Req packets, PTSPs, and PTSE
acknowledgment packets are exchanged.
6.3.7 LGN Horizontal Link Hello Protocol and PTSEs

The states of all the horizontal links to an LGN neighbor are determined from the information
in a single LGN Horizontal Link Extension information group included in the Hellos that are
sent over the SVCC-based RCC. Each distinct aggregation token value represents a distinct
horizontal link with its own independent state machine. This aggregation token is present in
the LGN Horizontal Link Extension IG. An LGN horizontal link is advertised in PTSEs origi-
nated by the LGN, if and only if, the LGN horizontal link hello state is twoWayInside. The
LGN Horizontal Link Hello Protocol is fully implemented as described in Sections 5.6.3 and
5.6.3.2 of the ATMF PNNI Specification.

ATM Forum PNNI
6.3.8 PGLs
6.3.8.1 Parent Nodes and Child Nodes
PGL capability is fully implemented as described in Sections 5.10.1 through 5.10.1.5 of the
ATMF PNNI Specification. A PGL is a node in a peer group that collects and assimilates data
to represent the entire peer group as a single node in the form of a parent node. A parent node
is an LGN that represents its peer group at the next higher level of hierarchy. A child node is a
node at the next lower level of hierarchy. It can be an LGN or a physical node.
6.3.8.2 PGL Election
ATM Forum PNNI

The PGL is elected by the members in the peer group. The node with the highest leadership
priority is chosen as PGL. After a node becomes PGL, its leadership priority is increased by 50
to ensure stability. The elections take place on a continuing basis such that when a node with a
higher leadership priority than the current PGL becomes active, another election is held to
transfer the leadership. Elections are also held when the current PGL fails or is deleted, so that
the leadership is transferred to the node with the next highest priority. If more than one node
has the same priority, the node with the highest node ID is declared PGL and its leadership
priority is increased by 50 to ensure stability.
6.3.8.3 PGL Capability

The PGL performs the following functions:
• The PGL provides support for the AwaitUnanimity, OperPGL, and
HungElection states of the PGL FSM. All of the other FSM states are already
implemented in ForeThought 5.x.
• The PGL allows the PGL priority to be set to non-zero values.
• The PGL allows the creation of parent-child relationships between nodes to facili-
tate LGN activation.
• The PGL provides support for the activation of a parent LGN when a child node
enters the OperPGL state in its PGL FSM.

ATM Forum PNNI
6.3.9 Aggregation Tokens and Link Aggregation

6.3.9.1 Aggregation Tokens
An aggregation token is a number assigned to an outside link by the border nodes at the ends of
the outside link. The same number is assigned to the uplink associated with the outside link.
In all parent peer groups and higher level peer groups, all uplinks that have the same aggrega-
tion token are combined, or aggregated, and represented as one link.
Aggregation tokens are fully implemented as described in Sections 5.10.1 through 5.10.1.5 of
the ATMF PNNI Specification. All links between a pair of LGNs with same aggregation token
must be advertised as one logical link. Derived uplinks and LGN horizontal links are aggre-
gated depending on the aggregation token values included in the PTSEs describing uplinks
and horizontal links as binding information.
6.3.9.2 Derived Aggregation Tokens

If neighbors exchange configured aggregation tokens of the same value, then the derived
aggregation token is that value (e.g., if both advertise 8, the derived value is 8; or if both
advertise 0, the derived value is 0).
If one neighbor has a configured aggregation token of 0 and the other has a non-zero config-
ured aggregation token, then the derived aggregation token is the non-zero value (e.g., if one
advertises 4 and the other advertises 0, then the derived value is 4).
If neighbors exchange configured aggregated tokens of different non-zero values, then the
derived aggregation token value is 0 (e.g., if one advertises 5 and the other advertises 6, then
the derived value is 0).
link x 9
LGN A link y 8 LGN B
link a 9
link b 9
S1 S2 link c 8 S3 S4
Peer Group A Peer Group B
Figure 6.4 - Aggregation Example

ATM Forum PNNI
In the aggregation example in Figure 6.4, switches S1 and S2 form peer group A and switches
S3 and S4 form peer group B. Outside links a and b each have an aggregation token of 9, and
outside link c has has an aggregation token of 8. Switches S2 and S3, which are border nodes,
advertise three separate uplinks. However, links a and b are aggregated together as one LGN
horizontal link, which is link x. There is another separate LGN horizontal link for link c, which
is represented as link y.
6.3.9.3 Link Aggregation

Link aggregation is fully implemented as described in Section 5.10.3.1 of the ATMF PNNI
Specification. Link aggregation is the process of representing several parallel links as a single
higher-level link. The topology state parameters for this type of link are derived from those for
ATM Forum PNNI

the individual links being aggregated.
For example, suppose you have two outside links with same aggregation token as follows:
• link1 (OC-12c on an LC network module) Non-VP capable
• link2 (OC-3c on a Series D network module) VP capable
The LGN horizontal link advertises the RAIGs of the link with highest bandwidth (available
cell rate) per service category and advertises it as VP capable. Even if only one of the aggre-
gated links is VP capable, then the aggregated link is considered to be VP capable.
Therefore, in the example links given, the aggregated link advertises OC-12 values for all ser-
vice categories, except for ABR, because ABR is not supported on LC network modules. The
OC-3’s values are advertised, instead, for ABR.
6.3.10 Nodal Aggregation

Nodal aggregation is implemented as described in Sections 5.10.3.2 and 5.10.3.3 of the ATMF
PNNI Specification, except for bypasses and exceptions. Nodal state parameter information
group is used to advertise elements of this node’s complex node representation. All metrics
and attributes for each input and output port must be included in the same IG. Additionally,
the flag in the nodal IG is set for complex nodal representation as specified in Section 5.14.9.1.2
of the ATMF PNNI Specification.
6.3.11 Internal and Exterior Address Summarization

Address summarization is fully implemented as described in Section 5.9 of the ATMF PNNI
Specification. When summarizing exterior reachable addresses, the aggregation of RAIGs
associated with the address being summarized is performed only if there is a summarizing
policy on that prefix with metrics (RAIGS) specified. Otherwise, no RAIG aggregation is per-
formed.

ATM Forum PNNI
6.3.12 Anycast Address Scoping

Scope defines to what level in the PNNI routing hierarchy an address should be advertised.
(See Section 6.4.3.3.2 in this manual for more information about scope.) Anycast address scop-
ing is implemented as described in Section 7.0, Annex 5, and Annex 6 in the UNI 4.0 Specifica-
tion; Section 9.4.2.4 in the ILMI 4.0 Specification; and in Sections 3.5.1, 5.3.6, and 5.9.1, and
Appendix A in the PNNI v1.0 Specification.
The handling of address scoping for anycast (group) addresses consists of two parts:
• the mapping of organizational scopes into routing levels
• the handling of the connection scope selection IE during connection establishment
6.3.12.1 Organizational Scope and PNNI Scope

The organizational scope of an anycast address is learned by the switch when the end device
registers the anycast address through ILMI registration. The organizational scope is mapped
into a routing level following a mapping table that is configured using the pnniScopeMap-
pingTable table in the PNNI MIB. The mapping occurs when the address is registered as
reachable through a PNNI node. Then this information is stored in Prefix Table Entries (PTEs)
and flooded to other nodes through PTSEs.
The PNNI scope mapping levels can be configured for each PNNI node through SNMP. If no
levels are configured, then default values are used. Similarly at ILMI registration, if no scope is
specified, then default values are used. The default values for organizational scope are local-
Network for ATM group addresses, and global for individual addresses.
6.3.12.2 Connection Scope Selection

The connection scope selection IE also gets mapped into a PNNI routing level. The mapping
takes place at the DTL originator. The PNNI routing level (PNNI node’s level) is used to select
PTEs that satisfy the scoping requirement. Only those that have been registered with the origina-
tor’s level that are numerically greater or equal to the mapped connection scope selection are eli-
gible for path selection. At the entry border node to the destination peer group, further checking
is performed. This time, the connection scope selection is not used. The scope of the highest level
DTL is used instead, and is compared to the advertisement scope of the PTEs.
There are two types of ATM endsystem addresses: individual and group addresses. An ATM
group has one or more members, and an ATM end system can be a member of zero or more
ATM groups at any time. The purpose of the connection scope selection IE is to allow the call-
ing user to indicate to the network the selection range/domain. A setup message with ATM
group address is received by the network. If the network does not find any registered
addresses or members with the indicated connection scope, then the network releases the call
with cause 3 (no route to destination). For ATM group addresses, if no connection selection
scope is specified in the setup, then routing uses localNetwork as the default to route to the
group address. For PNNI routing, unicast or anycast/group addresses are treated alike.

ATM Forum PNNI
6.4 Internetworking between PNNI and FT-PNNI

Some networks running FT-PNNI may be connected to networks running PNNI. To provide
internetworking of these protocols, ForeThought 5.x and 6.x allows reachability information to
be leaked dynamically between FT-PNNI and PNNI peer groups.
6.4.1 Gateway Switches and Split Switches

6.4.1.1 Using ForeThought 5.x
ATM Forum PNNI

When you upgrade a switch to ForeThought 5.x, it has a single FT-PNNI node by default. If you
change the default domain protocol to gateway, a second node is created automatically on the
same switch that runs the PNNI protocol, so you have both a single FT-PNNI node and a sin-
gle PNNI node running concurrently within a single switch. This is called a gateway switch. A
gateway switch connects a PNNI network with a FT-PNNI network.
Using ForeThought 5.x, you can also create a second PNNI node on single switch. A switch that
is running two or more PNNI nodes is called a split switch. A split switch connects two or more
PNNI areas.
The FT-PNNI and PNNI nodes within a gateway switch and the multiple nodes within a split
switch do not share topology information, but they dynamically exchange reachability infor-
mation to facilitate connectivity between a source in one network and a destination in another.
The process for this exchange of reachability information is discussed in Section 6.4.3.
As far as DTL processing is concerned, a gateway switch or a split switch appears as either a
DTL terminating or DTL originating node. For example, Figure 6.5 shows that S3 is a DTL ter-
minating node for the FT-PNNI DTL and a DTL originating node for the PNNI DTL for a con-
nection going from switch S1 to switch S4.
area
S3 gateway switch
PNNI node
S1
S3
FT-PNNI node
S2 S4
Figure 6.5 - Internetworking of FT-PNNI and PNNI

ATM Forum PNNI
6.4.1.2 Using ForeThought 6.x

When you upgrade a switch to ForeThought 6.x, it has a single PNNI node by default. Using
ForeThought 6.x, you can have up to five levels of hierarchy using PNNI. You can create a split
switch with a maximum of five lowest level nodes. All five nodes can be at the same forelevel
or four can be at the same forelevel with one at a higher (numerically smaller) forelevel. Each
of these nodes can have a PNNI instance, and each PNNI instance can have one lowest level
node and up to nine LGNs.
You can also create a gateway switch with a maximum of four lowest level PNNI nodes and
one lowest level FT-PNNI node. All five nodes can be at the same forelevel or four can be at
the same forelevel with one at a higher (numerically smaller) forelevel.
6.4.2 Interfaces
Interfaces can be attached to PNNI nodes. By default, a PNNI interface comes up attached to
the default PNNI node in ForeThought 6.x. FT-PNNI interfaces are automatically attached to
FT-PNNI nodes. An interface becomes a PNNI or FT-PNNI interface via ILMI auto-configura-
tion or by the user hard configuring the UNI using the command signalling new. (For more
information about this command, see Part 3 of the AMI Configuration Commands Reference
Manual.)
6.4.3 Dynamic Leaking of Reachability Information

This feature allows dynamic leaking of reachability information between the single FT-PNNI
node and the single PNNI node within a gateway switch. It may also be used to control the
leaking of reachability information between multiple PNNI peer groups. Using split switches,
it is possible to connect and exchange reachability information between any number of PNNI
peer groups. This dynamic leaking is controlled using the following user-configurable set-
tings: grouping of nodes within a switch into areas and domains, and configuring policies that
allow you to summarize, suppress, or advertise reachability addresses between peer groups.
See Section 6.4.3.3.1 for more information about policies. Nodes within a switch that belong to
different domains do not exchange information.

ATM Forum PNNI
6.4.3.1 Areas
An area is a subset of nodes within a domain that are contiguous and that together execute a
link state routing protocol to exchange reachability information dynamically among them-
selves. Because of the database exchange protocol, two nodes that belong to an area have iden-
tical copies of the link-state topology database. Using ForeThought 5.x, an area can be one of
the following:
• a non-hierarchical PNNI network (i.e., a single PNNI peer group because only the
lowest-level peer group is implemented)
• a hierarchical FT-PNNI network (i.e., a multiple FT-PNNI peer group system
using hierarchy)
ATM Forum PNNI

• a non-hierarchical FT-PNNI network (i.e., a single FT-PNNI peer group)
Using ForeThought 6.x, an area can be any of the above or can be the following:
• a hierarchical PNNI network (i.e., a multiple PNNI peer group system using hier-
archy)
Split switches can be used to connect multiple PNNI areas and gateway switches can be used
to connect an FT-PNNI area and a PNNI area, as shown in Figure 6.6. The only restriction in
configuring areas is that a FT-PNNI area cannot adjoin another FT-PNNI area.
PNNI node Switch S2

FT-PNNI node (split switch)
Area A
Area D
Area B
Switch S1 Area C Switch S3

(gateway switch) (gateway switch)
Figure 6.6 - Split Switches and Gateway Switches Connecting Areas

ATM Forum PNNI
6.4.3.1.1 Peer Groups in Areas

Multiple peer groups can exist within an area. Peer groups within an area are connected to
each other by border links as shown in Figure 6.7.
C.6
Peer Group Peer Group D
Border Link D.4
C.5
D.3 C.4
Switch/Node D.2 C.3 Peer Group C
Horizontal Link
D.1 C.2
C.1
A.1 B.2
B.3 B.8
A.2
B.4
A.3 B.1
B.5 B.7
Peer Group A
B.6
Peer Group B
Figure 6.7 - Peer Groups Connected by Border Links
6.4.3.1.2 Area IDs

Each area has an area ID which is unique only within a domain. The area ID is only useful at
the border of the area. Since area IDs are included in some of the routing packets distributed
throughout a domain, all switches belonging to the same area must be configured with the
same area ID. It is especially important for the nodes in split switches and gateway switches to
be configured with the correct area ID.

ATM Forum PNNI
6.4.3.1.3 Levels
Each area has a level associated with it. Levels are used to control the flow of reachability infor-
mation in a multi-level routing hierarchy (which is discussed in the next section). An N-level
hierarchy has N discrete values of levels associated with it, namely, L1, L2,..., LN. Each of the
areas in the hierarchy has one of these N values as its level. Numerical values are assigned to
levels in order from the lowest to the highest; i.e., L1 < L2 < L3 < LN, so LN is the highest level.
An area of level Li can adjoin any number of areas at level Li-1, but can only adjoin at most one
area at level Li+1 or higher. This means that on a given switch, there can never be areas config-
ured that belong to more than two distinct levels. Furthermore, there can only be one area that
belongs to the higher level.
ATM Forum PNNI

6.4.3.2 Domains
A domain is a group of areas that are configured to dynamically exchange reachability infor-
mation with one another. This allows connectivity between end systems belonging to different
areas without configuring static routes between areas. However, reachability information is
exchanged between domains only if inter-domain routes are manually configured between the
two domains. Figure 6.8 shows an example of how an inter-domain route can connect two
domains.
PNNI node
FT-PNNI node
IISP link
Switch S2
(split switch)
Area A Area D
Area B Area C
Switch S1 Switch S3
(gateway switch) Node 1 Node 2 (gateway switch)
Domain 1 Domain 2
Figure 6.8 - Inter-domain Route Connecting Two Domains

ATM Forum PNNI
In Figure 6.8, the two nodes configured in switch S1 belong to different areas, and so they do
not share the same link-state topology database. The same is true for the nodes configured in
switch S3. Dynamic reachability leaking takes place among the nodes in switch S1, enabling
end systems in Area A to reach other end systems in Area B, and vice versa within Domain 1.
The same applies to the nodes in switch S3.
The two nodes in split switch S2 belong to two different areas, each of which is part of a differ-
ent domain. An inter-domain route must be configured on both node 1 and node 2 so they can
share reachability information between two domains. An inter-domain route can be created
using the command routing pnni interdomain new. Exchange of reachability information
can then occur between the nodes over this inter-domain route.
6.4.3.2.1 Configuring Domains

Domains only need to be configured on split switches and gateway switches; i.e., switches
that belong to multiple areas. Other switches operate properly without domain configuration.
On switches running ForeThought 5.x, one FT-PNNI domain (with domain index 1) is pre-con-
figured by default. On switches running ForeThought 6.x, one PNNI domain (with domain
index 1) is pre-configured by default. The default domain cannot be deleted, but the informa-
tion within the domain can be changed. Multiple domains can be configured on a switch. For
each domain, the following parameters can be changed:
• The domain name, which is optional, can be configured for easier manageability.
However, the name of the default domain cannot be changed.
• The domain protocol, which sets the routing protocol for the switch, can be set to
ftpnni, pnni, or gateway. This parameter can also be changed on the default
domain. On switches running ForeThought 5.x, the default setting is ftpnni, so if
you want to configure a split switch, you must change the domain protocol to
pnni. On switches running ForeThought 6.x, the default setting is pnni, so if you
want to configure a split switch, you simply need to create one or more additional
PNNI nodes on the switch. On switches running ForeThought 5.x or 6.x, if you
want to configure a gateway switch, you must change the domain protocol to
gateway.
• The domain prefix can be changed. This sets the default 13-byte NSAP prefix used
by the switch’s routing protocol(s).
• A signalling interface (port/VPI) can be associated with a particular domain,
while the routing interface (port/VPI) for that signalling interface can be associ-
ated with a particular routing node within that domain.
For more information about changing these parameters, see Chapter 4 in Part 2 of the AMI

ATM Forum PNNI
6.4.3.3 Propagation of Reachability Information

The leaking of reachability information between two areas is constrained by two things: policy
and scope.
6.4.3.3.1 Policies
ForeThought software lets you configure a policy as a flexible means of enforcing security
across the network topology. When a node discovers an address that is leaked by another
node, it checks its own policy table to determine if the address is to be advertised within its
own area. There are three types of policies:
• summary - Addresses matching a summary policy cause just the summarized
ATM Forum PNNI

prefix to be announced to the node’s peer group instead of advertising each prefix
individually.
• suppress - Addresses matching a suppress policy are not announced to the
node’s peer group at all.
• advertise - Addresses matching an advertise policy cause the entire address to
be announced to the node’s peer group.
6.4.3.3.1.1 Policies in a Flat PNNI Network

In each node’s policy table, the switch chooses the best address match, meaning the longest
prefix match, to determine which policy applies to an address. Therefore, the suppress and
advertise policies provide limited filtering since addresses matching a broad prefix can be
filtered at the split switch or gateway switch through the suppress policy, then a particular
service, device, or switch can be exempted by advertising an address within the suppressed
range. The summary policy is best for scalability since one address prefix is shared, rather
than dozens or hundreds of more specific addresses.
For example, suppose you have an area A which contains peer groups A.1 and A.2. Suppose
you create two policies: one says suppress all address that are area A and the other policy says
advertise all address that are peer group A.1. If an address A.1 comes in to the area, the policy
for advertise takes precedence over the policy for suppress (the advertise policy is a longer
prefix match). Therefore, the address is advertised. However, if an address A.2 comes in to the
area, it is suppressed (since there is no advertise or summary match for A.2).
6.4.3.3.1.2 Policies in a Hierarchical PNNI Network

In an LGN hierarchy, when you assign a policy to an LGN, the information that is present in
the lower level peer group, or in the contained peer group, gets summarized, suppressed, or
advertised, depending on the type of policy configured. For example, suppose you have a
peer group B.1.1 which has a higher level LGN B.1. If there are exterior reachable addresses in
peer group B.1.1, you can apply a policy to LGN B.1 which suppresses or summarizes those
addresses.

ATM Forum PNNI
6.4.3.3.2 Scope
6.4.3.3.2.1 Scope in a Flat PNNI Network

Each piece of reachability information has a source area ID and a scope associated with it. The
source area ID of reachability information originated in a given area is 0 within that area,
meaning that is local. The source area ID of reachability information advertised from, for
example, area A1 into area A2 has source area ID A1.
When a leaked address is to be advertised or summarized, a node also determines the scope
with which to advertise the address (or its summary). The scope denotes the highest level in
the PNNI routing hierarchy at which an address can be advertised. Scoping ensures that
reachability information does not loop. A reachable address being advertised with a scope of l
can only be leaked to areas of level lower (larger in numerical value) than l. The scope of
reachability information originated in a given area is 0 within that area, meaning that it is
local. The scope of reachability information advertised from, for example, area A1 (at level l1)
into area A2 (at level l2) is l2 if l2 is a lower level than l1, but is l1 if l2 is a higher level than l1.
6.4.3.3.2.2 Scope in a Hierarchical PNNI Network

See Section 6.3.12 in this manual for information about scope in a hierarchical PNNI network.
6.4.3.3.3 The Process for Leaking Reachability Information

The following process is used by a node to determine if and how a leaked reachable address is
to be advertised to its peer group:
1. The node finds the longest matching policy prefix for the address.
2. If there is no matching policy prefix, the leaked address is advertised as is. If there
are multiple addresses with different scopes, then multiple addresses are
advertised.
3. If the longest matching policy prefix is a policy suppress, the leaked address is not
advertised at all.
4. If the longest matching policy prefix is a policy advertise or summary, the policy
prefix is advertised or summarized.
5. If there are multiple leaked addresses with different scopes for which the policy
prefix is longest matching, then the one with the widest scope is used to determine
the scope with which to advertise or summarize the policy prefix.
The following rule determines the scope with which to advertise an address or a policy prefix:
• If the address used to determine the scope (chosen above) was leaked from a
strictly lower level (numerically larger) area, then the scope of the address is used;
otherwise, the level of the area to which the advertising node belongs is used.
Metrics can optionally be configured with a policy prefix. In this case, the configured metrics
are included in the advertisement of the policy prefix.

ATM Forum PNNI
6.5 Path Extension QoS

This feature allows QoS parameters (CTD, CDV, and CLR) to be configured for VPs. For a
PNNI interface that is configured on a VPs, the associated QoS parameters are computed as the
combination of the parameters configured on the VP and those that are part of the switch and
network modules.
This feature also allows a particular service category to be configured as “none,” in which case
the associated PNNI link does not advertise the service category, thus preventing connections
with the given service category from using the link. For more information about assigning a
path extension to a VP, see Chapter 3 in Part 1 of the AMI Configuration Commands Reference
ATM Forum PNNI

Manual.

ATM Forum PNNI
6.6 Load-balanced Routing

As a value-added feature, you can perform load-balanced routing on FORE switches when
using ATMF PNNI. This feature lets you configure a node to distribute CBR, VBR, ABR, or
UBR traffic more evenly across available links. This lets you use your resources more effi-
ciently and helps to relieve congestion in the network.
ATMF PNNI path computation is done in two steps. In the first step, the shortest path is com-
puted using only links that satisfy the profile’s QOS requirements. For a load-balanced profile,
only one shortest path tree is computed, but up to four alternate links that lead to the equal
cost shortest paths are stored at each hop to be used in the second step. For a non load-bal-
anced profile, three shortest paths are computed and optimized for administrative weight,
CTD, and CDV.
In the second step, a DTL is constructed by tracing the path from the destination to the source.
For load-balanced profiles, up to four choices are available at each hop and one is selected ran-
domly, thereby distributing the calls across the network. However, because the links are
selected randomly, there is no guarantee of satisfying the connection’s CTD and CDV require-
ments. Therefore, when a load-balanced profile fails to meet the CTD and CDV requirements,
a non load-balancing profile is used to compute the path. If load-balancing is enabled on a
node, then any user-defined profile with load-balancing disabled is ignored. Similarly, if load-
balancing is disabled on a node, then any user-defined profile with load-balancing enabled is
ignored.
This feature is enabled (set to true) by default on a per-node basis, but it can be disabled (set to
false) during the creation of a node by entering routing pnni node new -index
<integer> -loadbalancing false. It can also be enabled or disabled on an existing node
using routing pnni node modify -index <integer> -loadbalancing (true|false).
For more information about these commands, see Chapter 4 in Part 2 of the AMI Configuration
Commands Reference Manual.

ATM Forum PNNI
6.7 Load-balanced (Congestion-based) UBR Routing

This feature lets the switch perform congestion-based, load-balanced routing for UBR SPVXs.
The switch uses existing parameters to measure the amount of congestion on the UBR links,
and then advertises that information to decide how to route a call. If multiple paths exist that
are equal cost, the switch tries to choose the path that is least congested.
Two important functions must occur to use congestion-based information in UBR routing.
First, the current congestion must be measured and advertised by all switches. Second, that
congestion information must be used by the source switch during path computation.
ATM Forum PNNI

Every call using congestion-based UBR routing has the following two additional parameters
associated with it:
• Minimum Cell Rate (MCR) - This value represents the minimum expected cell
rate to be used by the call. It is used when computing the congestion information
to be advertised to other switches.
• Peak Cell Rate (PCR) - The value represents the expected peak cell rate for the call.
It is used during path computation to insure that the call’s bandwidth require-
ments are met.
6.7.1 Advertising Congestion Metrics

Switches advertise two pieces of information about link congestion: the amount of link band-
width available for UBR traffic, and the amount of bandwidth used by UBR traffic. They are
computed as follows:
• Available UBR Bandwidth - This parameter is equal to the total link capacity,
minus the bandwidth reserved for other, higher priority service categories. For
example, if a link has a capacity of 100cps, and there is currently a CBR call of
40cps over that link, then the available UBR bandwidth is (100 - 40) = 60cps.
• Used UBR Bandwidth - This parameter is calculated by taking the maximum of
the following two parameters. The first one is the measured UBR traffic over that
link. The second is the total MCR from all congestion-based calls over the link.
For example, if the link mentioned previously has two congestion-based UBR calls
on it, with MCR values of 20cps and 10cps, the sum of all MCRs is 30cps. Then this
value is compared to the measured UBR traffic. If the measured traffic is 40cps, then
40cps is advertised since it is greater than 30cps. This algorithm ensures that con-
gestion is always “reserved” for the MCR values of all congestion-based UBR calls.
Both the available and used UBR cell rates are advertised in PNNI PTSEs, and flooded to the
rest of the network. In this way, every switch has a comprehensive view of the congestion of
the entire network.

ATM Forum PNNI
6.7.2 Routing Using Congestion Information

The idea of congestion-based routing is to route calls over links that have excess bandwidth
available for the call. To this end, path computation initially tries to route the call in con-
strained mode, in which only links with sufficient resources are considered. If no route can be
found to the destination in this mode, then unconstrained mode is used, in which all links are
considered regardless of the amount of congestion on them.
To determine if a link has enough bandwidth to be considered in constrained mode, the
unused UBR bandwidth on the link must be determined. This value is calculated by subtract-
ing the used UBR bandwidth from the available UBR bandwidth (these are the two values
advertised by the switch in the PTSEs). If the resulting value of unused UBR bandwidth does
not at least equal the PCR of the call, the link is not considered in constrained mode. Among
the paths that satisfy the PCR requirement of a call, the following parameters are used to
choose the most preferable path. These requirements are listed in order of priority.
• Path Cost - The least cost paths are always preferred.
• Maximum Congestion - Among equal cost paths, only paths that have a maxi-
mum congestion within a range of the least known congested path are consid-
ered. This range is called the PC congestion range, and can be configured using
using the -pccongesrange option under routing pnni node new. All paths
within the PC congestion range are considered equal. (See Section 6.7.2.)
• Average Congestion - Among equally congested paths, some preference is given
to those paths with the least average congestion. This factor only impacts routing
decisions if there are a large number of suitable links from a single node.
Only links that satisfy the PCR requirements of

NOTE the call use the previously listed criteria. If a
lower cost path does not satisfy the PCR
requirement, a higher cost path may be chosen
by congestion-based UBR routing.
If no paths are found to the destination using constrained routing, then path computation uses
unconstrained mode, in which only the administrative cost is considered.
Non-FORE switches and older FORE switches

NOTE do not advertise congestion information.
Congestion-based UBR routing will not consider
these links in constrained mode. Therefore, they
will not be used until all links advertised
congestion information are fully congested.

ATM Forum PNNI
ASX-1000, ASX-1200, and TNX-1100 backplane

NOTE links always advertise 0% congestion. This
ensures that multiple hops are never taken over
backplane links, which is not supported.
SPVXs are not rerouted based on congestion.

NOTE SPVXs are rerouted only if a lower cost path is
found.
6.7.3 Enabling Congestion-based UBR Routing
ATM Forum PNNI

There are some values that must be enabled to used congestion-based UBR routing for SPVXs
on a switch running ForeThought 6.1.x or greater.
• You need to enable the master congestion-based UBR routing flag using the
-lbubrenable option under routing pnni parameters modify. The default
value is false, which means that no SPVXs will use congestion-based routing,
even if the UPC contracts specify congestion-based parameters.
• You need to specify the -lbmincr and -lbpcr parameters under connections
upc new-ubr when configuring a new UBR contract to be applied to the SPVX.
6.7.4 Other Parameters Affecting Congestion-based UBR

There are other values that can affect congestion-based UBR routing:
• Available Cell Rate Proportional Multiplier (AVCRPM) and Available Cell Rate
Minimum Threshold (AVCRMT) - These parameters are the link significant
change thresholds. Usually, they are applied to the available cell rate advertised
on a link, but they are also used to check for significant changes in available UBR
bandwidth and used UBR bandwidth. If a link’s current values for either of these
parameters exceeds both the current value multiplied by the AVCRPM and the
link capacity multiplied by the AVCRMT, then the current values will be adver-
tised to the network. These parameters can be configured using the -avcrpm and
-avcrmt options, respectively, under routing pnni node modify or new.
• PC Congestion Range - This value is compared to the difference in congestion
between a candidate path and the lowest congested path during path computa-
tion (in constrained mode only). If the new path’s congestion exceeds the lowest
congested path by more than the PC congestion range, it is discarded and not con-
sidered when routing in constrained mode. The value can be configured using
using the -pccongesrange option under routing pnni node modify or new.

ATM Forum PNNI
• Profiles - A user-created PNNI profile can be configured as a congestion-based

profile. This means that the precomputed routes must consider the amount of
congestion. These profiles are only valid for congestion-based calls. To create a
congestion-based profile, use routing pnni profile new with the
-congestionbased option as true:
When using the -congestionbased option,

NOTE the -servicecategory must be UBR.
6.7.5 Load-balanced Routing vs. Congestion-based Routing

There are differences between standard load-balancing, as presented in Section 6.6, and con-
gestion-based load-balancing, as described in this section. Standard load-balancing randomly
chooses different links in a profile when generating a path in order to spread the load over
those paths. No attempt is made to prefer paths that are less loaded (congested). Standard
load-balancing is available for all service categories in ATMF PNNI.
In contrast, congestion-based load-balancing considers the amount of congestion on the links
and tries to spread the load evenly based on existing bandwidth. Additionally, this type of
load-balancing is only available for the UBR service category because all of the other service
categories are guaranteed to meet call requirements.

ATM Forum PNNI
6.8 Configuring PNNI

This section assumes that you are configuring PNNI for the first time. If you already have an
FT-PNNI network that you want to convert to PNNI, or if you already have a flat PNNI net-
work running and you want to change it to a hierarchical network, please see Appendix C in
this manual for examples of how to convert those networks.
Before you configure a PNNI network, it is assumed that you are familiar with the fundamen-
tals of PNNI routing and that you are familiar with FORE’s implementation of PNNI as
described in the earlier sections of this chapter.
ATM Forum PNNI

There are a few major steps that the system administrator should follow in order to configure
a PNNI network.
1. Do preparation work first to determine the following:
a. Decide how to divide the NSAP address space.
b. Decide how many layers of hierarchy you need.
2. Configure the NSAP addresses on all of the switches.
3. Choose a PGL.
4. Configure the parent nodes.
5. Configure the PGL priority.
6. Configure the remaining peer groups.
7. Verify that the hierarchy has come up.
The remainder of this section gives an example of configuring a PNNI network.
To configure hierarchical PNNI on a switch, you will log into AMI on a switch running
ForeThought 6.0.x or greater and use the commands found under routing pnni.
More information about each of these commands

NOTE may be found in Chapter 4 in Part 2 of the AMI
6.8.1 Doing the Preparation Work

Before you configure any switches, you need to determine what you want your network to
look like. It is recommended that you sketch the intended topology on paper first to determine
the best way to segment the switches into peer groups.

ATM Forum PNNI
6.8.1.1 Dividing the NSAP Address Space

Determining the NSAP addresses is an important task when configuring a PNNI network.
You need to assign a common prefix to each member in a peer group. In this way, each peer
group can be identified outside of the group as a single node by its address prefix. This con-
serves switch memory by reducing the size of the routing table.
There are three address formats: DCC, ICD, and E.164. FORE switches implement the ICD
ATM format. As per the UNI 3.0 specification, all private networks should accept initial call
set-up messages containing ATM addresses with any of the approved formats and forward
the calls as necessary.
An NSAP address consists of the following:
• a 13-byte network-side prefix - The prefix is the NSAP prefix of the switch to
which the host is attached.
• a seven-byte user-side part - This consists of the following:
- a six-byte End System Identifier (ESI) - The ESI is the unique IEEE MAC
address of the interface.
- a one-byte selector field
6.8.1.2 Address Summarization

A summarized address is represented by a full NSAP address plus a bit mask, which is
referred to as a level. The level indicates the summary prefix for the address. That is, the level
is the length of the bit mask at which the address prefix becomes significant to PNNI routing
computations. This facilitates hierarchical routing because only summarized topology infor-
mation is advertised outside of a peer group.
The 13-byte prefix consists of 104 bits. The peer group ID (PGID) is calculated as (104 - n)
rightmost bits set to 0, where n is the level.
6.8.1.3 Determining the Layers of Hierarchy Needed

Address assignment to a node is determined by its location in the network topology. This allo-
cation of bits should allow for topology growth both from the top and bottom of the hierarchy
to ensure maximum scalability and flexibility. In Section 6.8.1.3.1, an enterprise design exam-
ple is given. In Section 6.8.1.3.2, a service provider design example is given. In Section
6.8.1.3.3, a configuration example is given that is used throughout the rest of this chapter.

ATM Forum PNNI
6.8.1.3.1 Enterprise Example

In a 13-byte prefix, the first 7 bytes are fixed. The next 6 bytes can be used for hierarchical
addressing. The following is one example of how you can use your addressing space:
• Reserved - 1 byte
• Region - 1 byte
• Campus - 1 byte
• Building - 1 byte
• Closet - 1 byte
• Switch fabric - 1 byte
ATM Forum PNNI

The following is an example of an NSAP prefix in ICD format:
Table 6.2 - Example Enterprise NSAP Prefix
Fixed: Specified by the Registration Variable: Used for Hierarchical Addressing

Authority Scheme
Reserved Rsvd Region Campus Bldg Closet Fabric
1 2 3 4 5 6 7 8 9 10 11 12 13
47 00 05 80 FF E1 00 F2 15 XX XX XX XX

ATM Forum PNNI
6.8.1.3.2 Service Provider Example

In a 13-byte prefix, the first 7 bytes are fixed. The next 6 bytes can be used for hierarchical
addressing. The following is one example of how you can use your addressing space:
• Reserved - 1 byte
• Backbone - 1 byte allows for 256 BB nodes
• Region - 1 byte allows for 256 regions within the backbone
• Metropolitan area - 1 byte allows for 256 metropolitan areas within each region
• Customer - 2 bytes allows for 65,536 customers within each metropolitan area
The following is an example of an NSAP prefix in ICD format:
Table 6.3 - Example Service Provider NSAP Prefix
Fixed: Specified by the Registration Variable: Used for Hierarchical Addressing

Authority Scheme
Reserved Rsvd BB Region Metro Customer
1 2 3 4 5 6 7 8 9 10 11 12 and13
47 AB CD EF 12 34 56 XX XX XX XX XXXX

ATM Forum PNNI
6.8.1.3.3 Configuration Example

The configuration example in Figure 6.9 shows a small network with 13 switches. The
switches are going to be divided into three peer groups that are each at level 80 (the default for
FORE switches). These three peer groups are going to be summarized by LGNs in one peer
group at a higher level of 72. This is an example of a simple two-level hierarchy.
A.X.2
A.X.1 A.Z.1
A.Z.3
ATM Forum PNNI

A.X.4 A.Z.2
A.Y.2
A.X.3
A.Y.1 A.Z.4
A.Y.3
A.Y.4 A.Y.5
Figure 6.9 - Example Network to be Configured
6.8.2 Configuring the NSAP Addresses

Configure one peer group at a time. In the example shown in Figure 6.9, start with what will
become peer group A.X and begin with switch A.X.1.
1. Display the default protocol for switch A.X.1. Be sure that all switches are running
PNNI as shown in the Defproto field under routing domain show.
switch_ax1:routing-> domain show

ID Name Defproto Defsum Prefix
ID Name Defproto Defsum Prefix
1 default pnni disabled 47.000580ffe1000000f31a0172
DefaultPrefix: 47.000580ffe1000000f21a26a3
PNNI is the default protocol on switches running ForeThought 6.0.x and greater. If
any switches are running FT-PNNI, you need to change the default protocol to
PNNI using the routing domain modify command. You can change the proto-
col when you use this command to modify the prefix in step 3.

ATM Forum PNNI
2. Check the connections between the switches that will become peer group A.X (in
this case, 1B1 and 1D4) to be sure that the Interface field shows PNNI, the
SigVersion is pnni10, and the State is up.
switch_ax1:-> signalling show

Interface Type SigVersion
AtmIf VPI Oper Admin Oper Admin State ILMI Side RemoteAddress
1B1 0 PNNI auto pnni10 auto up up network 172.19.4.173
1D1 0 privUNI auto uni30 auto down down network
1D4 0 PNNI auto pnni10 auto up up network 172.19.13.219
1CTL 0 privUNI auto uni30 auto up down network
3. Modify the domain prefix of switch A.X.1 as follows:
You checked the default protocol in step 1. If the

NOTE default protocol currently is FT-PNNI, you
should type in -defaultproto pnni after the
-id when entering this command.
switch_ax1:-> routing domain modify

Usage:
[-id] <integer> ID
[[-defaultproto] <protocol>] Default Protocol
[[-prefix] <NSAP Prefix>] Prefix
[[-defsumstate] (enabled|disabled)] Default Summarization
switch_ax1:-> routing domain modify -id 1 -prefix

47.0005.80.ffe100.0000.f200.0172
This operation requires the switch to be rebooted.

Any mismatch between the configured domain protocol and the interface type of
an existing UNI will result in that UNI not being created after reboot.
Reply ‘n’ to discard/correct any changes made so far.
Reboot the switch [n]? y
4. Repeat step 1 through step 3 on each switch in peer group A.X and assign the same
prefix to the other switches in that peer group. That is, the bits in the prefix should
be the same for all switches in the peer group, up to the level being used. The bits
after the level being used, up to the 104th bit, must be unique for all peers in the
peer group.

ATM Forum PNNI
6.8.3 Choosing a PGL

You need to determine which switches should be PGL capable (eligible to be elected as PGL)
for each peer group. When determining the best switches to be PGL, select switches with sev-
eral links that are well connected to all of the other members of its peer group. It is best to
choose two or three nodes to be PGL capable in case the PGL becomes disabled (e.g., a link
gets broken). If the PGL goes down, another election is held, and a new PGL is elected.
When configuring a PGL at any level of hierarchy,

NOTE it is highly recommended that you use a switch
that has a 32MB SCP or greater to be the PGL.
ATM Forum PNNI

6.8.4 Configuring the Parent Nodes
In the example in Figure 6.9, the best choices for PGL in peer group A.X are A.X.2 and A.X.4.
1. Configure a parent node only for the switches in peer group A.X that should have
PGL capability. Begin with switch A.X.2 as follows:
switch_ax2:routing-> pnni node newparent

[-parent] <integer> Node
[-child] <integer> Node
[[-level] <integer>] Level
[[-pglprio] <integer>] Leadership Priority
switch_ax2:routing-> pnni node newparent -parent 2 -child 1
The default level is 80. When a parent node is created, the level is decremented by
8 by default. Do not assign a PGL priority yet. Create all of the parent nodes for the
peer group and then set the PGL priority to a non-zero value on the PGL
candidates.

ATM Forum PNNI
2. Use the following command to display information about the parent node:
switch_ax2:routing-> pnni node show -index 2

Node PnniNodeID
2 72:80:47.000580ffe1000000f3000000.ff1a2e760002.00
PnniAtmAddress
0x47.000580ffe1000000f31a0519.ff1a2e760002.00
PeerGroupID Level Lowest Rstrn
72:47.000580ffe100000000000000 72 false false
Ptses OperStat AdminStat Shutdown ForeLevel ForeArea
0 down up false 5 5
DomainId DomainName LoadBalance PcCongeRnge
1 default true 50%
The first part of the PnniNodeID is 72. This shows the level of this node (the parent
node). The second part of the PnniNodeID is 80. This shows the level of the child
node. The OperStat is down because the PGL has not been configured and
elected yet.
6.8.5 Configuring the PGL Priority

In the example in Figure 6.9, the best choices for PGL in peer group A.X are A.X.2 and A.X.4.
Switches A.X.2 and A.X.4 should be configured with a PGL priority (-pglprio) that is non-
zero, which means they are eligible for election as PGL. The range of values is from 0 - 205.
The other switches A.X.1 and A.X.3 already have the default PGL priority of 0. Leave their pri-
ority at 0 so they are not eligible to be elected.
1. Configure the PGL priority on the switches that you want to be eligible for election.
In this example, they are switches A.X.2 and A.X.4.
switch_ax2:routing-> pnni node modify -index 1 -pglprio 100
switch_ax4:routing-> pnni node modify -index 1 -pglprio 50

ATM Forum PNNI
2. The election process may take a minute or two. Display the PGL information for
the nodes on switch A.X.2 to see if the election is complete. If the PglState says
operPgl when the node has been elected as PGL.
switch_ax2:routing-> pnni node show -pgl

Node PglPrio(Cfg) AdvPglPrio ParentIx PglInitTime PglOverDelay PglReelect
(sec) (sec) (sec)
1 50 100 2 15 30 15
PglState PglTimeStamp
operPgl TUE JAN 26 19:24:28 1999
PreferredPgl
ATM Forum PNNI

80:160:47.000580ffe1000000f31a0519.ff1a2e760001.00
ActiveParentNodeId
72:80:47.000580ffe1000000f3000000.ff1a2e760002.00
CurrentPgl
80:160:47.000580ffe1000000f31a0519.ff1a2e760001.00
Node PglPrio(Cfg) AdvPglPrio ParentIx PglInitTime PglOverDelay PglReelect

(sec) (sec) (sec)
2 0 0 0 15 30 15
PglState PglTimeStamp
starting N/A
PreferredPgl
0:0:00.000000000000000000000000.000000000000.00
ActiveParentNodeId
0:0:00.000000000000000000000000.000000000000.00
CurrentPgl
0:0:00.000000000000000000000000.000000000000.00
The advertised PGL priority (AdvPglPrio)

NOTE shows the user-configured priority for all nodes
except the PGL. Because the priority of the PGL
is incremented by 50 to ensure stability, the
AdvPglPrio of the PGL shows the user-
configured priority plus 50.

ATM Forum PNNI
6.8.6 Configuring the Remaining Peer Groups

After the PGL has been elected for peer group A.X, the network looks like Figure 6.10 with
LGN A.X representing peer group A.X. LGN A.X is at level 72, while the rest of the lower level
nodes are at level 80.
A.X PNNI node

Level 72 PGL
LGN
PG A.X
Level 80 A.X.2
A.X.1 A.Z.1
A.Z.3
A.X.4 A.Z.2
A.Y.2
A.X.3
A.Y.1 A.Z.4
A.Y.3
A.Y.4 A.Y.5
Figure 6.10 - LGN A.X Represents Peer Group A.X at Level 72
Repeat the steps in Section 6.8.2 through Section 6.8.5 to configure peer groups A.Y and A.Z.
Once you have configured peer groups A.Y and A.Z, the network looks like Figure 6.11. LGN
A.Y is at level 72 representing peer group A.Y, while the rest of the lower level nodes are at
level 80. Similarly, LGN A.Z is at level 72 representing peer group A.Z, while the rest of the
lower level nodes are at level 80.
The LGNs form peer group A at level 72. The LGNs dynamically establish SVCC RCCs to each
other and begin to communicate.

ATM Forum PNNI
PG A
Level 72
A.X A.Y A.Z
ATM Forum PNNI

PG A.X
Level 80 A.X.2
A.X.1 A.Z.1
A.Z.3
A.X.4 A.Y.2 A.Z.2

A.X.3 A.Z.4
PG A.Z
PG A.Y Level 80
A.Y.1 A.Y.3 Level 80
A.Y.4 A.Y.5
Figure 6.11 - Completed Peer Groups

ATM Forum PNNI
6.8.7 Verifying That the Hierarchy Has Come Up

There are several things you can check to ensure that the hierarchy is working properly.
1. Check the HelloSt on the outside link between any two peer groups as follows:
switchaz2::routing-> show
AtmIf VPI Node Domain IntType SigProto SigSt
1C1 0 0 1 auto privNNI up
NodeST HelloSt PeerSt
up twoWayInside full
1C2 0 0 1 auto privUNI down
N/A N/A N/A
1C3 0 0 1 auto privNNI up
up commonOutside N/A
Press <return> for more, or 'q' to quit...q
When everything is configured properly, the HelloSt on the outside link between
two peer groups (border link) goes to commonOutside as shown on port 1C3. If
the link still shows twoWayOutside, then check the following items:
a. Ensure the prefixes have been properly configured using routing
domain show. Remember that each switch in a peer groups shares a pre-
fix, and each peer group must have a different prefix than the other peer
groups.
b. Ensure the parent nodes have been properly configured and that their
OperStat is up using routing pnni node show.
c. Ensure that the PGLs have been properly configured and that their
PglState is operPgl using routing pnni node show -pgl. The bor-
der link still shows twoWayOutside if the neighboring peer group does
not have a PGL configured yet or if the neighboring peer group’s PGL has
not come up yet.

ATM Forum PNNI
2. Check that an SVCC RCC has been established between LGNs by using the follow-
ing command on each LGN. If the SVCC RCC has been established properly, the
HelloState shows twoWayInside.
switchay2:-> routing pnni svccrcc

Node SvccRccIx Version HelloState RcvHellos XmtHellos
2 1 version1point0 twoWayInside 15 15
RemoteNodeId IfIndex
72:80:47.000580ffe1000000f2000000.ff150e130002.00 1245184
RemoteAtmAddress VPI VCI
0x47.000580ffe1000000f2150e13.ff150e130002.00 0 33
ATM Forum PNNI

switchaz2:-> routing pnni svccrcc
Node SvccRccIx Version HelloState RcvHellos XmtHellos
2 1 version1point0 twoWayInside 18 18
RemoteNodeId IfIndex
72:80:47.000580ffe1000000f3000000.ff1a2e760002.00 1179648
RemoteAtmAddress VPI VCI
47.000580ffe1000000f31a0519.ff1a2e760002.00 0 33
In ForeThought 6.0, policing is disabled by default

NOTE on NNI links. When configuring PNNI, it is
recommended that you leave the policing off and
that you disable the policing on NNI links on
any third-party switches that are part of the
network. If you enable policing on the NNI links,
the SVCC RCCs would be subject to policing and
cells could be dropped on the SVCC RCCs.

ATM Forum PNNI
3. Check that a horizontal link has been established between LGNs by using the fol-
lowing command on each LGN. If the horizontal link has been established
properly, the HelloState shows twoWayInside and the LinkType shows
horizontalLinkToFromLgn.
switchaz2::routing-> pnni link show -nodeindex 2

Node LinkIf PortId HelloState Version
1 1114112 30000010 twoWayInside version1point0
LinkType SvccRccIx
horizontalLinkToFromLgn 1
RemoteNodeId RmtPortId
72:80:47.000580ffe1000000f2000000.ff150e130002.00 10000013
DerAggrTok UpnodeId
0 0:0:0x0000000000000000000000000000000000000000
UpNodeAtmAddress CommonPGID
0x0000000000000000000000000000000000000000 0:00.000000000000000000000000

This chapter contains information that pertains to the signalling portion of ForeThought soft-
ware. This signalling information is described in the following sections:
• Section 7.1 - UNI 4.0 Supplementary Services
• Section 7.2 - Virtual UNI
• Section 7.3 - Proxy Signalling Services
• Section 7.4 - VCI Allocation Range
• Section 7.5 - Signalling Scope
• Section 7.6 - Signalling Channel Auto Configuration Procedures
• Section 7.7 - Allowable Combination of Traffic Parameters
7.1 UNI 4.0 Supplementary Services
Signalling
ForeThought 6.1 or greater supports supplementary UNI 4.0 services as described in the UNI
4.0 signalling specification. This implementation of supplementary services satisfies all of the
mandatory requirements, as specified in Annex 4 of the ATM Forum’s ATM UNI Signalling
Specification, Version 4.0.
The following supplementary services are supported by FORE ATM switches:
• Direct Dialing In (DDI)
• Multiple Subscriber Number (MSN)
• Calling Line Identification Presentation (CLIP)
• Calling Line Identification Restriction (CLIR)
• Connected Line Identification Presentation (COLP)
• Connected Line Identification Restriction (COLR)
• Sub Addressing (SUB)
• User to User Signalling (UUS)
The details of these supplementary services are described in the following sections.
For information on using the AMI commands for supporting different supplementary ser-
vices, see Section 7.1.2.

Signalling
7.1.1 Description of Supplementary Services

This section gives a high level description of the different supplementary services.
Figure 7.1 - UNI 4.0 Supplementary Services

Signalling
7.1.1.1 Direct Dialing In (DDI)

The DDI supplementary service enables a user to call directly to another user on the ISDN pri-
vate branch exchange (ISPBX) or other private system without human intervention. In this
context, a DDI number is a part of the ISDN number significant to an ISPBX or another private
system. DDI is provided when at least a part of the ISDN number significant to the called user
is passed to that user. These last digits are received by the ISPBX or another private system,
which automatically establishes a call to the destination without the assistance of an operator.
The DDI number is delivered from the network to the called user (in the called party number)
using the Q.2931 procedures. If the addressing/numbering plan identification is “ISDN/tele-
phony numbering plan (recommendation E.164),” the type of the number is included in the
called party number IE and is coded as “unknown or subscriber number, national or interna-
tional number.”
The insertion and subsequent removal of DDI

NOTE related information is a feature of an ISDN-
BISDN interworking device. The switch will not
support the insertion/removal of DDI and
related information since it is not an
interworking device. The switch is only capable
of carrying information between such
Signalling
interworking devices.
7.1.1.2 Multiple Subscriber Number (MSN)

The MSN supplementary service allows the assignment of multiple ISDN numbers to an inter-
face. This capability allows a calling user to select one or multiple distinct terminals or to iden-
tify the station that may provide other supplementary services. The Q.2931 call control
procedures are used at the originating network side. If the MSN supplementary service is pro-
vided to the subscriber, the network may use this information in the calling party number IE,
to identify the calling party.
The insertion and subsequent removal of MSN

NOTE related information is a feature of an ISDN-
BISDN interworking device. The switch will not
support the insertion/removal of MSN related
information since it is not an interworking
device. The switch is only capable of carrying
information between such interworking devices.

Signalling
7.1.1.3 Calling Line Identification Presentation (CLIP)

The CLIP supplementary service provides the calling party’s ISDN number to the called party,
possibly together with the subaddress information. The provided information consists of the
ISDN number of the calling user in a form sufficient to allow the called user to return the call
(i.e., a subscriber number, a national number, or an international number and optionally, a
subaddress if provided by the calling user).
If ILMI is enabled on the interface, the calling party number can be derived from the ILMI
address registration procedures. If ILMI is disabled on the interface, then the default calling
party address (presently configurable) can be used as the calling party address.
7.1.1.4 Calling Line Identification Restriction (CLIR)

The CLIR service is the reverse of the CLIP service. CLIR is a supplementary service offered to
the calling party to restrict the presentation of the calling party’s ISDN number and subad-
dress to the called party. When a CLIR service is applicable and activated, the originating net-
work provides the destination network with an indication that the calling user’s ISDN
number and subaddress (if provided by the calling user) are not allowed to be presented to the
called user and the corresponding IEs are not included in the SETUP message sent to the
called party.
7.1.1.5 Connected Line Identification Presentation (COLP)

COLP is a supplementary service offered to the calling party. The connected party number is
used to provide OSI X.213 services. This is also used to identify the connected party in case of
call transfer, PBX-to-PBX connections. This service is used to provide the connected party’s
ISDN number with possible additional subaddress information to the calling party. This sup-
plementary service is not used to check the dialing, instead used as an indication to the calling
subscriber of the connected ISDN number. All information pertaining to the COLP is inserted
in the CONNECT message sent as a part of the Q.2931 call procedures. If no information is
provided by the connected user, the network provides the default number associated with this
user’s access at the destination switch. The COLP supplementary service makes use of the
connected number and connected subaddress IEs inserted in the CONNECT message.
If ILMI is enabled the connected address can be derived from the ILMI address registration
procedures. If ILMI is disabled, the connected address should be made configurable for each
signalling interface.

Signalling
7.1.1.6 Connected Line Identification restriction (COLR)

COLR is a supplementary service offered to the connected party to restrict presentation of the
connected party’s ISDN number and subaddress to the calling party at the UNI. When COLR
is applicable and activated, the destination network provides the originating network with an
indication the connected user’s ISDN number and subaddress information are not allowed to
be presented to the calling user. In this case, no connected number and subaddress informa-
tion is provided to the calling user in the CONNECT message.
You can restrict the presentation of the connected number to the calling user.
7.1.1.7 Sub Addressing (SUB)

The SUB supplementary service allows the called user to expand his or her addressing capac-
ity beyond the one given by the B-ISDN number at the UNI.
7.1.1.8 User-to-User Signalling (UUS)

The UUS supplementary service (specified in Q.2957) allows a B-ISDN user to send/receive a
limited amount of information to/from another B-ISDN user over the signalling virtual chan-
nel in association with the call/connection to the other B-ISDN user.
For example, the communicating entities may exchange some authentication protocol while
the call is being established. The UUS data is up to 128 bytes and is used for compatibility
Signalling
checking between the called and calling user.
This is required for applications that are supporting OSI X.213 services and X.25. It is also used
between PBX-to-PBX to pass additional information for the call.

Signalling
7.1.2 Configuring Supplementary Services

The options to enable or disable the supplementary services are under the signalling new
command. You can use the -suppservices option to globally enable or disable the services
for the signalling interface. Supplementary services is disabled by default.
The pertinent portion of the AMI syntax is listed here for reference:
myswitch:signalling-> new
[-vpi] <unsigned> VPI
[[-suppservices] (enabled|disabled)] Suppl Services (default: disabled)
[[-clip] (enabled|disabled)] CLIP Status (default: disabled)
[[-clir] (enabled|disabled)] CLIR Status (default: disabled)
[[-colp] (enabled|disabled)] COLP Status (default: disabled)
[[-colr] (enabled|disabled)] COLR Status (default: disabled)
[[-sub] (enabled|disabled)] SUB Status (default: disabled)
[[-uus] (enabled|disabled)] UUS Status (default: disabled)
[[-defconndpaddr] <NSAP Address>] Connected Party Def Addr
Supplementary services cannot be enabled for

NOTE unsupported signalling interfaces.
The -defconndpaddr option is used for both

NOTE CLIP service and also for address filtering.
Configuration of supplementary services for

NOTE interfaces of type PublicUNI and UNI40 is not
supported. Supported interfaces are
PrivateUNI/UNI40 and IISP /UNI40 only.

Signalling
To display supplementary services, enter the following:
myswitch:signalling-> show -services

Suppl
Port VPI Services CLIP CLIR COLP COLR SUB UUS
1A1 0 enabled disabled disabled disabled disabled disabled disabled
1A2 0 N/A N/A N/A N/A N/A N/A N/A
1A3 0 enabled enabled enabled enabled disabled enabled enabled
CLIP_addr=47.300000000000000000000178.ff1a3bf00001.00
COLP_addr=47.234567ff0000000000000178.ff1a3bf00001.00
1A4 0 disabled enabled disabled disabled enabled disabled enabled
CLIP_addr=47.300000000000000000000178.ff1b4ce40001.00
1B1 0 N/A N/A N/A N/A N/A N/A N/A
1B2 0 disabled disabled disabled disabled disabled enabled disabled
COLP_addr=47.234567ff0000000000000178.fd323cf00302.00
1B3 10 N/A N/A N/A N/A N/A N/A N/A
1B4 100 disabled disabled disabled disabled disabled disabled disabled
1C1 0 disabled disabled disabled disabled disabled disabled disabled
1C2 0 N/A N/A N/A N/A N/A N/A N/A
Signalling
1D1 0 N/A N/A N/A N/A N/A N/A N/A
1D2 0 N/A N/A N/A N/A N/A N/A N/A
1D3 0 enabled enabled disabled disabled disabled enabled enabled
CLIP_addr=47.234567ff0000000000000178.fd323cf34312.00
COLP_addr=47.234567ff0000000000000178.fd323cf34312.00
1CTL 0 N/A N/A N/A N/A N/A N/A N/A
To modify supplementary services, you would use the following command (pertinent portion
of the AMI syntax is listed here for reference):
myswitch:signalling-> modify
Usage:
[-vpi] <unsigned> VPI
[[-clip] (enabled|disabled)] CLIP Status
[[-clir] (enabled|disabled)] CLIR Status
[[-colp] (enabled|disabled)] COLP Status
[[-colr] (enabled|disabled)] COLR Status
[[-sub] (enabled|disabled)] SUB Status
[[-uus] (enabled|disabled)] UUS Status
[[-defconndpaddr] <NSAP Address>] Connected Party Def Addr

Signalling
7.2 Virtual UNI

Virtual UNIs are supported across a physical interface by assigning each end user, one or more
virtual path connections on its own interface and then using a VPCI group index identifying
the VPCI-VPI-port mapping for each of the virtual UNI signalling interfaces on that physical
interface.
One of the applications for virtual UNIs is in the DSL (Digital Subscriber Line) environment.
DSL devices can have a dedicated signalling channel and dedicated virtual path to an ATM
switch. As illustrated in Figure 7.2, DSL-to-ATM networks typically have a few DSL devices,
aggregated by a DSLAM. A DSLAM (Digital Subscriber Line Access Multiplexer) is a device
which takes all user VPCs and combines these onto one or more connections to an ATM switch
by relaying user cells based only on their VPI. In other words, a DSLAM is a VP-Mux devoid
of signalling interfaces.
DSLAM
(VP-Mux)
DSL ATM Cloud

devices Switch
Figure 7.2 - DSL-to-ATM Network Connection
In most cases, virtual UNIs have VCs allocated on VP=0. This means that all end users (DSL
devices) are signalling and requesting SVCs on VP=0. The VP-Mux maps each of the end users
to a unique VP on the switch side. Virtual UNI support requires the switch to understand that
setup messages received on a non-zero signalling path must be responded to as though all the
devices were on path zero. The ATM switch allocates virtual circuits on the VPs on which it
receives the Setup message. However, it translates the VPCI-VPI=0 when reporting it back to
the end user in the Connect message.
In order for the switch to support VP-Muxes, it supports multiple virtual UNIs on the single
physical interface to the VP-Mux. A unique VPI on the interface between the DSLAM and the
switch is used to distinguish between individual end users. Switches send and receive signal-
ling messages on a unique VPI associated with the individual user devices.

Signalling
The switch maintains a translation table between VPCI and VPI. For the end user, the VPCI is
equal to the VPI. The VPCI remains constant in the signalling message as it travels between
the user and switch. The VPI associated with the VPCI is used while the switch allocates a vir-
tual circuit upon receiving a Setup message. Similarly, the VPCI associated with the VPI on
which the circuit is allocated is used while sending the Connect message back to the end user.
The switch maintains two tables:
• VPCI mapping table (VMT): contains information for a single VPCI-VPI-port
mapping where each entry in the VMT is assigned a unique map index.
• VPCI mapping group table: contains a listing of all VPCI map indexes where each
entry in the VPCI group table is assigned a unique group index. This group index
must be used when creating the signalling interfaces on the switch. This enables
VPCI-VPI translation in the signalling messages.
The following outlines the steps necessary to create virtual UNI connections:
1. Create one or more VPCI-to-VPI mappings using the following AMI command:
signalling vpci maps new -mapindex -vpci -atmif -vpi
2. Create a VPCI group for one or more of the VPCI-to-VPI mappings created in the
previous step using the following AMI command:
Signalling
signalling vpci groups new -groupindex -position -mapindex
3. Create the signalling channel and specify the assigned VPCI group index created
in the previous step using the following AMI command:
signalling new
For usage parameters descriptions for the AMI commands mentioned above, see Part 3 of the
AMI Configuration Commands Reference Manual.

Signalling
The following example shows how a single VPCI-VPI mapping can be achieved for three vir-
tual UNIs using the VPCI mapping table (VMT). In Figure 7.3, a switch is connected to the
VP-Mux and three end user devices are connected to the VP-Mux.
User A
VPI=0
User B
Port 3
VPI=0 Switch
Port 4 4C8
Port 10
User C VPI=50 Port 5
VP-Mux
Port 3, VP=0 Port 10, VP=11
Figure 7.3 - Virtual UNI Network Configuration
The connection setup procedure for the above example is as follows:

1. User A, attached to port 3 of the VP-Mux, sends a setup request out on VPI=0. A
specific VCI is not required since the switch allocates any available VCI.
2. The VP-Mux performs ingress ports-to-VP mapping on egress port 10 and for-
wards the message to the switch on VPI=11, VCI=5.
3. The switch determines that VCI=53 is available within VPI=11 and assigns VPI=11,
VCI=53 to the end user.
4. The switch maps VPI=11 to VPI=0 and sends a Connect message informing the end
user to allocate a connection on VPI=0, VCI=53.
5. The VP-Mux performs VP-to-port mapping and forwards the message out port 3
over VPI=0, VCI=5 to the end user.
6. User A receives the Connect message and opens a circuit on VPI=0, VPI=53.

Signalling
The steps for configuring the switch for the virtual UNI example illustrated in Figure 7.3 is as
follows:
1. Create the VPCI-VPI-port mappings for the VPCI map table by entering the fol-
lowing AMI commands:
signalling vpci maps new 101 -vpci 0 -atmif 4C8 -vpi 11
You can simplify the above configuration by entering the required parameter values and tak-
ing the default values for the optional parameters. For example:
signalling vpci maps new 104
signalling vpci maps new 105 -vpci 50
2. Display the VPCI-VPI-port mappings that you just created by entering the
following AMI command:
Signalling
signalling vpci maps show
Map Index VPCI AtmIf VPI

101 0 4C8 11
102 0 4C8 12
103 50 4C8 13
3. Create a VPCI group for the VPCI-VPI-port mappings by entering the following
AMI commands:
signalling vpci groups new 1001 -mapindex 101
If you entered the simplified configuration, you would enter the following instead:
signalling vpci vpcigroup new 1004 1-mapindex 04
signalling vpcitable vpcigroup new 1005 -mapindex 105

Signalling
4. Display the VPCI group you just created by entering the following:
signalling vpci groups show
Group Index Position Map Index

1001 1 101
1002 1 102
1003 1 103
5. Create the signalling channels by entering the following AMI commands:
Make sure a VPT exists for that path before

NOTE creating the signalling channel (connections
path term new).
signalling new 4C8 11 -vpcigroup 1001
6. Display the VPCI group to which the signalling interface is associated by entering
the following:
signalling show -atm
AtmIf VPI SigVCI ILMIVCI InSigUpc OutSigUpc VpciGroupId

4c8 11 5 16 0 N/A 1001
4c8 12 5 16 0 N/A 1002
4c8 13 5 16 0 N/A 1003
For detailed information on the virtual UNI

NOTE commands, see Chapter 3 in Part 3 of the AMI

Signalling
The following must be taken into consideration when configuring virtual UNIs:
• There is a maximum configurable limit on the number of privates UNIs that can
be created. The maximum limit depends on the switch platform: HA/CF proces-
sors allow a maximum of 192 signalling interfaces and P5/P6 processors allow a
maximum of 600 signalling interfaces.
• A large number of NNI signalling interfaces is not addressed by virtual UNI sup-
port.
• Disable all debug tracing on the switch while using large number of private UNI
signalling interfaces.
• The default control port VC space is 0 to 1024 for the ASX-200BX, ASX-1000, 0 to
8192 for the ASX-4000, and 0 to 4091 for the ESX-3000. Ensure that the VC space is
sufficient to support the number of UNIs on the switch. If it is not sufficient, mod-
ify the maximum VCI number to be supported on the control port using the
connections path term modify AMI command. See Chapter 3 in Part 1 of the
• VC space on the switch: By default, 511 VCCs are allocated per VP and in turn to
each signalling interface. For a large number of UNIs, it is recommended to con-
figure VPs with a smaller VC range between 0-40.
• If end systems can be configured to run signalling and ILMI protocol (5 and 16,
Signalling
respectively) on non-reserved VCCs, these can be configured on, for example,
VCC 32 and 33, respectively. In this case, the VC range per VP can be set between
32-40 with 6 VCCs used for other applications on the end systems.

Signalling
7.3 Proxy Signalling

In addition to the existing two methods of establishing end-to-end circuits: provisioning of
PVCs (including SPVCs), and signalling by the end station to create a new circuit on demand
(SVCs), ForeThought 6.1.x supports a third method, called proxy signalling. Proxy signalling
allows an entity (known as the Proxy Signalling Agent or PSA) to perform signalling for one or
more ATM end stations that are not capable of ATM signalling. The PSA may have one or more
signalling interfaces where each signalling interface will have its own ILMI channel. There
should be a proxy signalling interface on the ATM switch associated with each signalling inter-
face on the PSA. Each proxy signalling interface on the switch controls a group of virtual path
terminations.
This implementation of proxy signalling satisfies all of the mandatory requirements, as speci-
fied in Annex 2 of the ATM Forum’s ATM UNI Signalling Specification, Version 4.0.
An ATM network using proxy signalling consists of the following components:
• Proxy Clients: Proxy clients are ATM end stations incapable of ATM signalling.
• Proxy Signalling Agent (PSA): Performs proxy signalling on behalf of the proxy
clients. This is connected to an ATM cloud across a UNI. There are no restrictions
on the location of the PSA as long as the PSA’s signalling terminates on the proxy
signalling service. The proxy signalling agent can be remotely attached to the
switch. This attachment can be accomplished by either provisioning Permanent
Virtual Circuits (PVCs) or Smart PermanentVirtual Circuits (SPVCs)
• Proxy Signalling Service: The signalling service which when configured on an
ATM switch, will allow proxy clients to be directly attached to the switch. The
proxy signalling service must be co-located on the switch.
Proxy
Signalling
ATM
Services
ATM Switch
PSA 4b1 4d1
4c1 4c2 4c3 4c4
Proxy Proxy Proxy Proxy

Client Client Client Client
1 2 3 4
Figure 7.4 - Example of a Proxy Signalling Configuration

Signalling
In the above example, proxy clients 1, 2, 3, and 4 are connected to an ATM switch on 4C1,4C2,
4C3, and 4C4, respectively. The PSA is connected to the switch on 4B1. The VC setup messages
are exchanged on 4B1, but the.actual VCs are set on ports 4C1, 4C2, 4C3, and 4C4.
ForeThought 6.1.x allows the creation of multiple

NOTE proxy signalling service interfaces on the switch.
This lets multiple proxy signalling agents to be
attached to the same switch. In this way,
different proxy clients can be controlled by
different proxy signalling agents.
Creation of a proxy service signalling interface requires the creation of valid entries in the
VPCI mapping table and VPCI group table.
The following outlines the steps necessary to configure proxy signalling services:
1. Create the virtual path terminators that will be controlled by the proxy service sig-
nalling interface.
connections path term new 4C1 60
Signalling
2. Create VPCI -VPI-port mappings (make sure to create a mapping for the signalling
interface*).
signalling vpci maps new -mapindex 1 -vpci 0 -atmif 4C1 -vpi 60
* signalling vpci maps new -mapindex 8 -vpci 90

Signalling
3. Display the VPCI-VPI-port mappings you have created:
signalling vpci maps show
Map Index VPCI AtmIf VPI

1 0 4C1 60
2 20 4C2 70
3 40 4C3 80
4 60 4C4 90
8 90 THIS_ATMIF THIS_VPI
4. Create a VPCI group for the VPCI-VPI-port mapping and add map indices to that
group:
signalling vpci groups add 100 2-4:8
5. Display the VPCI groups:
signalling vpci groups show
Group Index Position Map Index

100 1 1
100 2 2
100 3 3
100 4 4
100 5 8
6. Create the proxy signalling service interface (if the signalling interface already
exists, delete and re-create it):
signalling new 4C1 0 -admintype privUNI -adminversion uni40 -sigscope

proxy -sigmode nonassoc -vpcigroup 100

Signalling
7. Check that the proxy signalling interface has been created:
signalling show -protocol
Scope Mode QoS Accelerated

AtmIf VPI Oper Admin Oper Admin ExpIndex Clear
4B2 0 VP auto nonAssoc auto 0 disabled
4C1 0 Proxy Proxy nonAssoc nonAssoc 0 disabled
4C2 0 VP auto nonAssoc auto 0 disabled
4D1 0 VP auto nonAssoc auto 0 disabled
4E1 0 VP auto nonAssoc auto 0 disabled
4E2 0 VP auto nonAssoc auto 0 disabled
4CTL 0 VP auto nonAssoc auto 0 disabled
Signalling

Signalling
7.4 VCI Allocation Range

The VCI allocation range is the range of VCI values in a path within which a signalling chan-
nel will allocate VCs. This range defaults to the VCI-range of the virtual path containing the
signalling channel (also referred to as the containing path).
If the signalling channel is a link-scope signalling

NOTE channel, then this range applies to the containing
path and all other paths within which the
signalling channel has allocated connections.
A VCI allocation range can be specified when a signalling channel is created using the ATM
Management Interface (AMI) command signalling new. The range is computed differently
depending on whether or not you have ILMI running.
7.4.1 Determining the VCI Allocation Range with ILMI Down

If ILMI is not operational and if the VCI allocation range of the signalling channel is specified,
then the range is determined as an intersection of the following:
• the VCI-range of the path containing the signalling channel
• the range specified for the signalling channel
For example, if you create a signalling channel with ILMI down on a VP that has no VCIs
reserved, and you specify a VCI range:
signalling new 1A3 65 -ilmiadmstat down -adminminvci 70 -adminmaxvci

200
the actual range is computed as an intersection based on the range you enter (70 - 200) and the
range available in that VP (1 - 511).

Signalling
To display what the actual VCI range is, enter the following command:
signalling show -atmranges
Admin Admin Oper Oper Admin Max Max

AtmIf VPI MinVCI MaxVCI MinVCI MaxVCI MaxVPI SvccVPI SvpcVPI SvpCapable
1A1 0 32 511 32 511 1022 0 0 N/A
1A2 0 32 511 32 511 1022 0 0 N/A
1A3 65 70 200 70 200 510 65 65 N/A
1A4 0 32 511 32 511 1022 0 0 N/A
1CTL 0 32 1023 32 1023 0 0 0 N/A
In this display, the range that you have requested is shown in the Admin MinVCI and Admin
MaxVCI fields. The range that is computed and allowed is displayed in the Oper MinVCI and
Oper MaxVCI fields. In this example, no VCIs are reserved on VP 65, so the requested range of
70 - 200 is allowed.
In another example, if there are VCIs already reserved on VP 65, as shown below in the
MinVCI and MaxVCI fields:
connections path term show
Signalling
1A1 0 term N/A 37.3K 1 511 35 pvc
1A1 0 orig N/A 0.8K 1 511 7 pvc
1A2 0 term N/A 0.8K 1 511 6 pvc
1A2 0 orig N/A 0.8K 1 511 6 pvc
1A3 65 term N/A 0.8K 1 120 6 pvc
1A3 65 orig N/A 0.8K 1 120 6 pvc
1A4 0 term N/A 0.8K 1 511 6 pvc
1A4 0 orig N/A 0.8K 1 511 6 pvc
1CTL 0 term N/A 45.4K 1 1023 85 pvc
1CTL 0 orig N/A 45.4K 1 1023 33 pvc
and you enter the following:
signalling new 1A3 65 -ilmiadmstat down -adminminvci 70 -adminmaxvci

200
then the actual range is computed as an intersection based on the range you enter (70 - 200)
and the range available in that VP (1 - 120).

Signalling
To see what the actual VCI range is, enter the following command:
Admin Admin Oper Oper Admin Oper Oper

Port VPI MinVCI MaxVCI MinVCI MaxVCI MaxVPI MaxSvccVPI MaxSvpcVPI SvpCapable
1A1 0 32 511 32 511 1022 N/A N/A N/A
1A2 0 32 511 32 511 1022 N/A N/A N/A
1A3 0 32 511 32 511 1022 N/A N/A N/A
1A3 65 70 200 70 120 1022 N/A N/A N/A
1A4 0 32 511 32 511 1022 N/A N/A N/A
1CTL 0 32 1023 32 1023 0 N/A N/A N/A
In this example, since VCIs 1 - 120 are already reserved on VP 65, the range of 70 - 120 is the
actual range allowed.
7.4.2 Determining the VCI Allocation Range with ILMI Up

If ILMI is operational, then the peer switch’s VCI allocation range is obtained by reading its
ILMI MIB variable atmfAtmLayerMaxVccs. The VCI allocation range of the signalling chan-
nel is then determined as an intersection of the following:
• the VCI-range of the path containing the signalling channel
• the range specified for the signalling channel (if applicable)
• the VCI-range of the peer signalling entity
For example, if your peer has no VCIs reserved, and you create a signalling channel with ILMI
up on a VP that has no VCIs reserved, and you specify a VCI range:
signalling new 1A3 65 -ilmiadmstat up -adminminvci 70 -adminmaxvci

200
the actual range is computed as an intersection based on the range you enter (70 - 200), the
range available in that VP (1 - 511), and the range of the peer signalling entity (32 - 511).

Signalling
To display what the actual VCI range is, enter the following command:

1A1 0 32 511 32 511 1022 N/A N/A N/A
1A2 0 32 511 32 511 1022 N/A N/A N/A
1A3 0 32 511 32 511 1022 N/A N/A N/A
1A3 65 70 200 70 200 1022 N/A N/A N/A
1A4 0 32 511 32 511 1022 N/A N/A N/A
1CTL 0 32 1023 32 1023 0 N/A N/A N/A
In this display, the range that you have requested is shown in the Admin MinVCI and Admin
MaxVCI fields. The range that is computed and allowed is displayed in the Oper MinVCI and
Oper MaxVCI fields. In this example, no VCIs are reserved on VP 65 and the signalling peer
supports the range of 32 - 511, so the requested range of 70 - 200 is allowed.
In another example, if the peer supports the range of 32 - 511, but there are VCIs already
reserved on VP 65, as shown below in the MinVCI and MaxVCI fields:
Signalling
connection path term show
Input Output
Port VPI Port VPI ResBW CurBW MinVCI MaxVCI VCs Protocol
1A1 0 terminate N/A 37.3K 1 511 35 pvc
1CTL 0 terminate N/A 45.4K 1 1023 85 pvc
originate 1A1 0 N/A 1.7K 1 511 7 pvc
Press return for more, q to quit: q
and you enter the following:

200
then the actual range is computed as an intersection based on the range you enter (70 - 200),
the range available in that VP (1 - 120), and the range supported by the peer (32 - 511).

Signalling

1A1 0 32 511 32 511 1022 N/A N/A N/A
1A2 0 32 511 32 511 1022 N/A N/A N/A
1A3 0 32 511 32 511 1022 N/A N/A N/A
1A3 65 70 200 70 120 1022 N/A N/A N/A
1A4 0 32 511 32 511 1022 N/A N/A N/A
1CTL 0 32 1023 32 1023 0 N/A N/A N/A
In this example, since the peer supports VCIs 32 - 511, but VCIs 1 -120 are already reserved on
VP 65, the range of 70 - 120 is the actual range allowed.
In a third example, if the peer only supports the range of 32 - 180, and VCIs 1 - 120 are already
reserved on VP 65, and you enter the following:

200
then the actual range is computed as an intersection based on the range you enter (70 - 200),
the range available in that VP (1 - 120), and the range supported by the peer (32 - 180).

1A1 0 32 511 32 511 1022 N/A N/A N/A
1A2 0 32 511 32 511 1022 N/A N/A N/A
1A3 0 32 511 32 511 1022 N/A N/A N/A
1A3 65 70 200 70 120 1022 N/A N/A N/A
1A4 0 32 511 32 511 1022 N/A N/A N/A
1CTL 0 32 1023 32 1023 0 N/A N/A N/A
In this example, since the peer supports VCIs 32 - 180, and VCIs 1 - 120 are already reserved on
VP 65, the range of 70 - 120 is the actual range allowed.

Signalling
7.5 Signalling Scope

A signalling channel can be configured to control either VCs within its own VP (VP-scope, the
default) or to control VCs in its own VP, in addition to controlling VCs in other VPs on the link
(link-scope).
The following criteria must be satisfied in order to use link-scope signalling channels:
• There can be only one link-scope signalling channel on a given link.
• A link-scope signalling channel must be created within VP 0 on the given link.
• The path containing the link-scope signalling channel must be an elastic path.
• A backplane signalling channel in an ASX-1000, ASX-1200, or a TNX-1100 cannot
be a link-scope signalling channel.
7.5.1 VC-Space
The VC-space of a signalling channel is the VPI/VCI space within which a signalling channel
can allocate connections. This is determined by the allocation scope of the signalling channel
and the VCI allocation range (see Section 7.4) for the signalling channel.
The VC-space of a VP-scope signalling channel is the VCI allocation range of the signalling
Signalling
channel applied only to the path that contains the signalling channel.
The VC-space of a link-scope signalling channel is comprised of the aggregate VCI allocation
range of the signalling channel applied to all paths controlled by the signalling channel (i.e.,
the path containing the signalling channel and all other paths in which the signalling channel
has allocated connections).

Signalling
7.5.2 Dynamic Paths

A link-scope signalling channel allocates new virtual channel connections based on the avail-
ability of virtual channel connections within the VC-space of the link-scope signalling channel
at that time. A link-scope signalling channel allocates connections in a new originating/termi-
nating path (called a dynamic path, because signalling uses such a path on the fly) when its cur-
rent VC-space is exhausted.
Every time a link-scope signalling channel uses a

NOTE new path for allocating connections, its VC-space
increases.
To avoid conflicts of ownership of the path, a link-scope signalling channel never uses an
existing provisioned terminating/originating path (apart from the path that contains the sig-
nalling channel) to allocate virtual channel connections.
Dynamic paths inherit the characteristics of the virtual path that contains the signalling chan-
nel (called the containing path). Therefore, dynamic paths have the same VCI-range as the con-
taining path and are elastic paths.
To see which paths are dynamic paths, look at the Protocol field associated with each path
under the AMI command connections path term show. This field displays q2931 for
dynamic paths and displays pvc for provisioned terminating/originating paths.
A dynamic path is controlled only by the link-scope signalling channel that uses it. Therefore,
a dynamic path is not open to any provisioning such as the creation of a PVC connection,
SPANS signalling path, etc.

Signalling
7.6 Signalling Channel Auto Configuration Procedures

This section provides an overview of signalling channel auto configuration and describes
some rules for configuring signalling interfaces. Auto configuration requires that ILMI is up
on that signalling interface.
7.6.1 Overview of Signalling Channel Auto Configuration

Signalling support for UNI 3.0, UNI 3.1, and UNI 4.0 is provided in ForeThought 5.2.x and
greater. The switch can be configured to detect which interfaces are UNI 3.0, which are UNI
3.1, and which are UNI 4.0 automatically. The switch also translates signalling messages
between interfaces of different types.
When a signalling channel is created using signalling new -adminversion auto, then
the creation is delayed until a response is obtained from the peer or until the ILMI keep alive
requests time out. Based on the peer response, the appropriate SSCOP stack is initialized. If
the peer does not support the atmfAtmLayerUniVersion variable, then the switch software
defaults to UNI 3.0. Similarly, if the ILMI keep alive requests time out, then the switch soft-
ware defaults to UNI 3.0. The switch periodically checks the version supported on the peer by
querying the atmfAtmLayerUniVersion variable. If the peer version has changed since the
last query, the signalling stack is recreated to reflect the change.
Signalling

Signalling
The following two tables summarize the above information. Table 7.1 highlights the actions
taken based on the configurations of the FORE switches on both sides of an interface.
ForeThought versions prior to ForeThought 4.1.x always default to UNI 3.0. Both UNI 3.1 and
UNI 4.0 use SSCOP version 31.
Table 7.1 - Action Taken Based on Both Switches’ Signalling Channel Configurations
Peer’s
Configured
Configured Action Taken
Version
Version
UNI 3.0 UNI 3.0 SSCOP 30 stack initialized
UNI 3.1 SSCOP 30 stack initialized; error message displayed on syslog
PNNI 1.0 SSCOP 30 stack initialized; error message displayed on syslog
Auto SSCOP 30 stack initialized; peer changes version to UNI 3.0
UNI 3.1 UNI 3.0 SSCOP 31 stack initialized; error message displayed on syslog
UNI 3.1 SSCOP 31 stack initialized
UNI4.0 SSCOP 31 stack initialized; error message displayed on syslog
PNNI 1.0 SSCOP 31 stack initialized; misconfiguration
PNNI 1.0 UNI 3.0 SSCOP 31 stack initialized; error message displayed on syslog
UNI 3.1 SSCOP 31 stack initialized; misconfiguration
UNI4.0 SSCOP 31 stack initialized; misconfiguration
PNNI 1.0 SSCOP 31 stack initialized
Auto SSCOP 31 stack initialized; peer changes version to PNNI 1.0
if it supports PNNI
UNI 4.0 UNI 3.0 SSCOP 31 stack initialized; error message displayed on syslog
UNI4.0 SSCOP 31 stack initialized
PNNI 1.0 SSCOP 31 stack initialized; misconfiguration
Auto UNI 3.0 SSCOP 30 stack initialized
UNI 3.1 SSCOP 31 stack initialized
UNI4.0 SSCOP 31 stack initialized
PNNI 1.0 SSCOP 31 stack initialized
Both stacks initialized as SSCOP 31 if both sides support UNI
Auto 3.1 or UNI 4.0; otherwise, both sides are initialized to SSCOP
30. If both sides support PNNI then, both sides are initialized
to SSCOP 31 and the version is set to PNNI 1.0.

Signalling
An alternative way of looking at the auto configuration procedure is to look at the

atmfAtmLayerUniVersion variable. Table 7.2 shows actions taken based on the
atmfAtmLayerUniVersion variable supported by the peer switch.
Table 7.2 - Action Taken Based on the Peer’s Supported MIB Variable
Configured atmfAtmLayer
Action Taken
UNI Version UNI Version
UNI 3.0 Version 3.0 SSCOP 30 stack initialized
Version 3.1 SSCOP 30 stack initialized; error message displayed on syslog
Version 2.0 SSCOP 30 stack initialized; version 2.0 is not supported and
error message displayed on syslog
MIB variable SSCOP 30 stack initialized
not supported
UNI 3.1 Version 3.0 SSCOP 31 stack initialized; error message displayed on syslog
Version 3.1 SSCOP 31 stack initialized
Signalling
MIB variable SSCOP 31 stack initialized; error message displayed on syslog
not supported
UNI 4.0 Version 3.0 SSCOP 30 stack initialized
not supported
Auto Version 3.0 SSCOP 30 stack initialized
not supported

Signalling
7.6.2 Rules for Signalling Channel Auto Configuration

In ForeThought 5.0.x and greater, there are two rules for signalling channel auto configuration.
The first deals with specifying the signalling interface type and signalling interface version.
The second rule concerns specifying the signalling scope and mode.
7.6.2.1 Specifying the Type and Interface Version

In AMI, the signalling interface type and signalling interface version are now treated as a pair.
The type and version either both have to be auto or both have to be given a parameter other
than auto. Specifying the type as auto and hard configuring the version or vice-versa is not
allowed. If a type is not entered, then the type parameter defaults to auto. Similarly, the ver-
sion parameter defaults to auto if a version is not entered.
The exceptions to this rule (to allow backward compatibility) is for the type to be entered as
publicUNI or IISP and the version to be entered as auto. Table 7.3 shows the valid combi-
nations.
Table 7.3 - Valid Type and Version Combinations
Type Version
auto auto
privateUNI uni30
privateUNI uni31
privateUNI uni40
publicUNI auto1
publicUNI uni30
publicUNI uni31
publicUNI uni40
IISP uni30
IISP uni31
IISP uni40
IISP auto1
privateNNI pnni10
1.
Exceptions to the rule.

Signalling
7.6.2.1.1 Examples of Valid Configurations

The following are examples of valid configurations to enter into AMI:
1. sig new 1a1 0

2. sig new 1a1 0 -admintype privateUNI -adminversion uni30
5. sig new 1a1 0 -admintype publicUNI
6. sig new 1a1 0 -admintype publicUNI -adminversion auto
7. sig new 1a1 0 -admintype publicUNI -adminversion uni30
10. sig new 1a1 0 -admintype IISP -adminversion uni30
13. sig new 1a1 0 -admintype privateNNI -adminversion pnni10
Signalling
The first example is the default that is most often used when configuring a signalling interface.
Using this specification allows the interface to auto configure to any possible valid combina-
tion like privUNI/uni30, privUNI/uni31, privUNI/uni40, ftpnni/uni30, ftpnni/
uni31, ftpnni/uni40, or privateNNI/pnni10. This is the most preferred input.
The fifth and sixth examples illustrate the exception to the rule.

Signalling
7.6.2.1.2 Examples of Invalid Configurations

Table 7.4 shows examples of invalid configurations to enter into AMI and the reasons why
they are invalid.
Table 7.4 - Invalid Type and Version Combinations
Invalid Combination Reason

sig new 1a1 0 -adminversion uni30 The type is not entered (auto).
sig new 1a1 0 -admintype auto -adminversion uni30 The type is auto.
sig new 1a1 0 -adminversion pnni10 The type is not entered (auto).
sig new 1a1 0 -admintype auto -adminversion pnni10 The type is auto.
sig new 1a1 0 -admintype privateUNI The version is not entered (auto).
sig new 1a1 0 -admintype privateUNI -adminversion The version is auto.
auto
sig new 1a1 0 -admintype privateUNI -adminversion pnni10 is an NNI protocol.
pnni10
sig new 1a1 0 -admintype publicUNI -adminversion pnni10 is an NNI protocol.
pnni10
sig new 1a1 0 -admintype IISP -adminversion IISP and pnni10 are conflicting NNI
pnni10 protocols.
sig new 1a1 0 -admintype privateNNI The version is not entered (auto).
sig new 1a1 0 -admintype privateNNI -adminver- The version is auto.
sion auto
sig new 1a1 0 -admintype privateNNI -adminver- The NNI type and the UNI version are
sion uni30 conflicting protocols.

Signalling
Table 7.4 - Invalid Type and Version Combinations (Continued)

7.6.2.2 Specifying the Scope and Mode

In AMI, the signalling scope (designated in AMI as -sigscope) and the signalling mode
(designated in AMI as -sigmode) are now treated as a pair. The pertinent portion of the AMI
syntax is listed here for reference:
myswitch:signalling-> new
Usage:
[[-sigmode] <alloc_mode>] Alloc Mode (default: auto)
[[-sigscope] <alloc_scope>] Alloc Scope (default: auto)
The scope and mode either both have to be auto or both have to be given a parameter other
than auto. Specifying the scope as auto and hard configuring the mode or vice-versa is not
allowed. If a scope is not entered, then the scope parameter defaults to auto. Similarly, the
Signalling
mode parameter defaults to auto if a mode is not entered.
Since scope and mode are not negotiated based on the peer signalling entity’s status, if the
scope and mode are hard configured on one side of a link, they should also be hard configured
on the remote side.
Table 7.5 shows the valid combinations.
Table 7.5 - Valid Scope and Mode Combinations
Scope Mode
autoScope autoMode
vpScope vpAssoc
vpScope nonAssoc1
linkScope nonAssoc
1.
This combination is valid as long as the -admintype is
not configured as auto or privateNNI. If the
-admintype is auto, the link could become a PNNI link,
which would conflict with a -sigmode of nonAssoc and
a -sigscope of vpScope.

Signalling
7.6.2.2.1 Examples of Valid Configurations

The following are examples of valid configurations to enter into AMI:
sig new 1a1 0

sig new 1a1 0 -sigscope auto -sigmode auto
sig new 1a1 0 -sigscope vp -sigmode vpAssoc
sig new 1a1 0 -sigscope link -sigmode nonAssoc
The first example is the default that is most often used when configuring a signalling interface.
This is the most preferred mode since the auto configuration of the link and scope procedures
can correctly determine what is valid for a given interface.

Signalling
7.6.2.2.2 Examples of Invalid Configurations

Table 7.6 shows examples of invalid configurations to enter into AMI and the reasons why
they are invalid.
Table 7.6 - Invalid Scope and Mode Combinations

sig new 1a1 0 -sigmode vpAssoc The scope is not entered (auto).
sig new 1a1 0 -sigscope auto -sigmode vpAssoc The scope is auto.
sig new 1a1 0 -sigmode nonAssoc The scope is not entered (auto).
sig new 1a1 0 -sigscope auto -sigmode nonAssoc The scope is auto.
sig new 1a1 0 -sigscope vp The mode is not entered (auto).
sig new 1a1 0 -sigscope vp -sigmode auto The mode is auto.
sig new 1a1 0 -sigscope link The mode is not entered (auto).
sig new 1a1 0 -sigscope link -sigmode auto The mode is auto.
sig new 1a1 0 -sigscope link -sigmode vpAssoc This is an invalid combination
Signalling
for an interface.
sig new 1a1 0 -sigscope vp -sigmode nonAssoc This is an invalid combination
if the type is auto (default).
sig new 1a1 0 -admintype privateNNI -adminversion VP scope and nonAssoc mode
pnni10 -sigscope vp -sigmode nonAssoc are incompatible with the type
privateNNI.
There is a potential invalid configuration that can occur because the auto configuration of the
operating scope and mode depend on whether a path is an elastic path or not. A non-elastic
path has bandwidth reserved for it from the link bandwidth at the time that the path is created
(when the -allocbw option in connections path term new is used.) An elastic path does
not have reserved bandwidth at creation time. Instead, it uses the residual bandwidth avail-
able on the link.
The invalid configuration can occur when a signalling interface on VPI 0 that has both the
scope and mode configured as auto becomes a PNNI interface (either by the user explicitly
specifying -admintype privateNNI or by specifying -admintype auto). In this case, if
the path is an elastic path, then the operating scope becomes link and the operating mode
becomes nonAssoc. If the path is a non-elastic path, then the operating scope becomes vp and
the operating mode becomes vpAssoc.

Signalling
Therefore, if a signalling interface exists between two switches on VPI 0, with one half of this
path being elastic on one switch and one half being non-elastic on the other switch, then the
interface comes up with incompatible modes on either side if it becomes a PNNI interface.
This results in no calls being set up on this interface.
To prevent this invalid configuration, be sure that both sides of your PNNI interface are either
elastic or non-elastic. If this invalid configuration occurs, simply delete one side of the path
and recreate it to match the other side.
7.6.2.3 Specifying the User/Network Side

While ForeThought automatically detects the side depending on the type, version, and routing
domain, there is one exception in the case of publicUNI. Unless the side is hard configured as
network, a publicUNI will always come up as user. To avoid this, configure the network
side first and then the user side will be automatically detected.

Signalling
7.7 Allowable Combination of Traffic Parameters
7.7.1 PNNI 1.0/UNI 4.0

The following information applies to both signalled connections and to PNNI SPVCs.
7.7.1.1 Service Categories

The ATM service categories assigned to connections in PNNI 1.0 and UNI 4.0 are defined in
the ATM Traffic Management Specification, Version 4.0 (TM 4.0). For PNNI 1.0 interfaces,
ForeThought 5.2.x and greater supports the service categories defined in TM 4.0 with a few
exceptions.
An explicit way of requesting a particular ATM service category is not provided in PNNI 1.0
and UNI 4.0 signalling. Instead, it must be derived from three pieces of information in the
SETUP message.
• the broadband bearer class in octet 5 of the broadband bearer capability (BBC)
information element
• the absence or presence of the ATM transfer capability (ATC) octet (octet 5a) in the
BBC information element
Signalling
• the value of the ATC, if present, and the absence or presence of the best effort indi-
cator in octet 18 of the ATM Traffic Descriptor information element
Derivation of a service category from the above information is specified in Section A9.2
(Determination of ATM service Category) of the ATM User-Network Interface (UNI)
Signalling Specification, Version 4.0 (UNI 4.0). However, ForeThought 5.2.x and greater does
not support the following from Table A9-1 of this specification:
• The Transparent VP-Service in the BBC information element is not supported.

Signalling
7.7.1.2 Allowable Combination of Traffic Parameters

The parameters specified in the BBC information element, the ATM traffic descriptor informa-
tion element, the Extended QoS parameter information element, the End-to-End transit delay
information element, and the QoS parameter information element of the SETUP message
should be consistent.
Table A9-2 of UNI 4.0 shows the allowable combinations. The following exceptions apply to
ForeThought 5.2.x and greater:
• In the Extended QoS parameters information element, for forward and backward
Cell Loss Ratio (CLR), ForeThought 5.2.x and greater does not distinguish between
the (CLP=0) and CLP (0+1) traffic streams.
• The PNNI 1.0 and UNI 4.0 specifications allow incomplete traffic contracts for
VBR in four instances (two for real time VBR and two for non-real time VBR).
Check note 7 and 8 of Table A9-2 of the UNI 4.0 specification. ForeThought 5.2.x
and greater does not support these four combinations.
7.7.2 UNI 3.X

FORE Systems supports a subset of traffic contracts which are specified in Table F.1 of the UNI
3.1 specification. These traffic contracts are enforced on both signalled connections as well on
PNNI SPVCs. The following are the allowable contracts supported by FORE’s switch fabrics.
• CBR: PCR (0), PCR (0+1)
• CBR: PCR (0+1)
• VBR: PCR (0+1), SCR, MBS (0)
• VBR: PCR (0+1),SCR, MBS (0+1)
• UBR: PCR (0+1)
Traffic contracts that do not completely

NOTE characterize the traffic characteristics of a given
service category are incomplete and not
supported. (For example, a VBR contract that is
received with an ATM traffic descriptor
information element that does not specify SCR
and MBS cell rates does not fully characterize a
VBR contract. Such a contract is, therefore,
rejected by the switch fabric).

Signalling
Table 7.7 shows the subset of the UNI 3.1 allowable combination of traffic parameters sup-
ported by ForeThought 5.2.x and greater:
Table 7.7 - UNI 3.1 Allowable Combination of Traffic Parameters in ForeThought 5.2.x and Greater
Broadband Bearer Class A X C X C X A X C X

Traffic Type CBR & & CBR &
Timing Required Y && && Y &&
PCR (CLP=0) S S
PCR (CLP=0+1) S S S S S S S S S S
SCR and MBS (CLP=0) S S
SCR and MBS (CLP=0+1) S S
Best Effort S S
Tagging Y/N Y/N Y/N Y/N N N N N N N
QoS Class * * * * * * * * 0 0
Signalling
The following are used in the table above:
Y Yes
N No
S Specified
Y/N Either Yes or No is allowed.
* All QoS classes are supported.
& The parameter is coded to either “no indication,” or
“VBR,” or, for UNI 3.1, Octet 5a (Traffic Type /
Timing Required) is absent; these 3 codings are
treated as equivalent.
&& The parameter is coded to either “no indication,” or
“No,” or, for UNI 3.1, Octet 5a (Traffic Type / Timing
Required) is absent; these 3 codings are treated as
equivalent.
A blank entry indicates that the parameter is not
present.

Signalling

CHAPTER 8 Security
ForeThought software version 5.0.x or greater provides various forms of switch security. Secu-
rity can be implemented by creating userids with different levels of privileges. The user login
authentication can be implemented using Kerberos and SecurID. There are also security meth-
ods that prevent access to the switch, which include IP filtering and NSAP filtering.
This chapter provides detailed information about each switch security method available. Steps
for configuring these methods are also provided in each section.
Before allowing access to a switch, the user must first be authenticated using one of the meth-
ods described in this chapter.
8.1 Configuring Userids

A user must first be authenticated to be allowed access to the switch. Once authenticated, the
user is given an authorization level. The network administrator creates and assigns a userid for
each user or for a group of users via the security login new command. See Part 2 of the
AMI Configuration Commands Reference Manual for more information about this command. A
userid consists of the following:
• a method of login authentication
• a level of AMI command privileges
• a level of AMI access levels
• an optional password
Security
On a new switch running ForeThought 6.1.x or greater, there is a default userid: ami. This user-
name is configured with the local authentication method, with admin privileges (meaning
you are allowed to use all AMI commands), and is allowed to log in to the switch using all the
possible methods. The userid is assigned a null password.
The network administrator should configure userids for all people who will have access to
each switch. If the administrator wants to use the same set of userids on each switch in the net-
work, he or she should configure the entire set of userids on each switch. This can also be done
with a batch file.

Security
The following AMI command is used to configure a new userid:
myswitch:security login-> new

Usage:
[-username] <text> User Name
[-application] <text> Application requiring authentication
[-authmethod] <text> Authentication Method
[[-profile] <text>] Profile Name (default: user)
The security login show command displays a list of user login methods:
myswitch:security login-> show

UserName Application Authentication Method Profile Name
--------------------- ----------- --------------------- --------------------
ami console password admin
ami http password admin
ami telnet password admin
private snmp community admin
public snmp community readonly
If a user attempting to access the switch matches one of the user login methods, then they are
permitted onto the switch. For example, if a user named ami telnets into AMI and supplies the
correct password, they are then permitted into the switch and given admin level privileges
(allowed to use all AMI commands). See Section 8.1.2 for more information.
Not all combinations of application and authentication methods are permitted. The following
table lists the allowable combinations:
Table 8.1 - Allowable Login Method Combinations
Application Method Authentication Method

telnet password, securid, kerberos
console password, securid
http password, securid
snmp community

Security
For some security methods, you can configure default login privileges so that you do not have
to add an entry to the security login table (security login show). Use the security
login defaults new command to add a default login privilege.
myswitch:security login defaults-> new

Usage:
[-authmethod] <text> Authentication Method
[-profile] <text> Profile Name
myswitch:security login defaults-> new kerberos admin

myswitch:security login defaults-> new securid admin
myswitch:security login defaults-> show

Authentication Method Profile Name
--------------------- --------------------
kerberos admin
securid admin
If there is a matching entry in the security login table, then that entry is used to determine the
profile. Otherwise, the profile is determined from the defaults table.
There are several different scenarios that can occur, such as the SecurID server being down or
a userid not being listed on a particular switch. Table 8.2 shows the different login scenarios
and the action taken by the switch for each.
Table 8.2 - Possible Login Scenarios
Login Scenarios Action
Security
Userid not listed and SecurID server not accessible Reject
Userid not listed; SecurID server accessible; and SecurID Reject
server rejects userid
Userid not listed; SecurID server accessible; and SecurID Accept
server accepts userid
Userid listed and access method not OK. Reject (If this is for SecurID or
Kerberos authentication, this
login will not be rejected if there
is a listing in the default table.)
Userid listed; access method OK; local authentication; and sup- Reject
plied password does not match

Security
Table 8.2 - Possible Login Scenarios (Continued)
Login Scenarios Action

Userid listed; access method OK; local authentication; and Accept
supplied password matches
Userid listed; access method OK; SecurID authentication; Reject
SecurID server accessible; and SecurID server rejects userid
Userid listed; access method OK; SecurID authentication; Accept
SecurID server accessible; and SecurID server accepts userid
SecurID server not accessible; and no local password
SecurID server not accessible; local password; and supplied
password does not match
Userid listed; access method OK; SecurID authentication; Accept
SecurID server not accessible; local password; and supplied
password matches

Security
8.1.1 Login Authentication Method

The network administrator can configure four different forms of login authentication: local
authentication, SecurID authentication, Kerberos authentication, and SNMP authentication.
The administrator may employ these methods for all users, or choose to employ the different
methods for different users.
8.1.1.1 Local Authentication

When a user is configured for local password authentication, he or she is prompted for a login
ID (userid) and a password which is stored locally in the switch whenever he or she attempts
to open an AMI session either via telnet or via the serial port. (The login password is not
required if the user tries to open an AMI session via a remote switch or via ForeView.) After a
validation check is made based on the scenarios listed in Table 8.2, an AMI session is started
(provided that a local AMI session is not already running).
Telnet from a FT 6.1 switch to another FT 6.1

NOTE switch requires password or SecurID
authentication. Prior to FT 6.1, remote AMI and
ForeView require SNMP authentication.
For specific information and examples of how to log in to the switch via telnet, the serial port,
or a remote switch, see the ATM Management Interface (AMI) Manual. For information about
logging in via ForeView, see the ForeView Network Management User’s Manual.
8.1.1.2 SecurID Authentication

When a user is configured for SecurID authentication and he or she attempts to log in, the user
is prompted for a login ID (userid) and password. The user should enter a SecurID passcode
when prompted for the password. The two-part passcode consists of: a secret, memorized per-
Security
sonal identification number (PIN) and the current code generated by the user’s assigned
SecurID token. After a validation check is made based on the scenarios listed in Table 8.2, an
AMI or HTTP session is started.
8.1.1.2.1 SecurID Protection on Switches

Security Dynamics ACE/server and client software1 prevents a user from logging into a
switch until the passcode entered by the user has been validated using an external security
server. The server uses SecurID tokens to validate the identity of users, and allows access only
to authorized users on valid clients (switches).
1.
The client software is already provided on FORE ATM switches, but the server software
must be purchased separately from Security Dynamics.

Security
Because SecurID does not protect SNMP or ILMI

NOTE access to the switch, it is recommended that
users either employ IP filtering as a selective
mechanism to allow SNMP changes, or disable
SETs from the network entirely.
8.1.1.2.2 SecurID Passcode

This authentication method provides a high level of security because the SecurID passcode
that allows access to the protected switches is comprised of two parts:
• a secret, memorized personal identification number (PIN)
• the current code generated by the user’s assigned SecurID token
8.1.1.2.2.1 PIN Number

The PIN is known only by the user. It can be either alpha-numeric or strictly numerical, and
can be either a fixed or variable length from 4 - 8 characters, depending on how the system
administrator configures the server.
8.1.1.2.2.2 SecurID Tokens

The second part is a unique code from the SecurID token that only the user possesses and
which cannot be counterfeited. Each authorized user on a protected system is assigned a
SecurID token to use when accessing a protected switch. SecurID tokens are small, hand-held
devices that use microprocessors to calculate and display random codes. These codes change
at a specified interval, which is usually once every minute. The random code displayed on a
user’s token is the same code the server software has generated for that moment.
8.1.1.2.3 SecurID Server

The server can run on a UNIX system or on a Windows NT system. Each FORE switch has a
defined set of authorized users. (See Part 2 of the AMI Configuration Commands Reference
Manual for more information.) When a user is designated as someone who has SecurID autho-
rization, that user is not permitted access to that switch until his or her identity is validated
based on the scenarios listed in Table 8.2.
When a user attempts to login with a SecurID passcode, the SecurID software running on the
switch verifies the passcode with the server; verifies the authenticity of the server so that no
other machine can pretend to be the server in order to capture security data; and encrypts and
decrypts messages sent between the switch and the server.
8.1.1.2.3.1 Slave Server

A backup, or slave server can be installed to ensure that authentication services are not inter-
rupted, even if the server goes down. When the slave detects that the master has failed, the
slave takes over authentication services. This failover is transparent to the user.

Security
8.1.1.2.3.2 Server Database

The server’s database includes records for all tokens, a list of switches to be protected, an
audit trail of SecurID and administrative activity, and a list of users who are authorized to
access each switch.
8.1.1.2.3.3 Data Encryption between the Server and Switches

Messages sent between the server and the switches are encrypted either using the DES algo-
rithm or Security Dynamics proprietary encryption algorithm. The server can use either
method, but all switches must use the same algorithm as the one configured on the server.
This data encryption method protects communications between the server and the switches
because the first time the switch contacts the server, it receives a node secret file, which is a
string of about 16 bytes. This string, which is known only to the server and this switch, is used
in encrypting messages between the server and the switch. Additionally, communications
between the master and slave servers and stored token information are encrypted. No one,
including system administrators, can breach security by inspecting secured token data.
8.1.1.2.4 SecurID AMI Commands

The security securid AMI command allows SecureID configuration on the switch. The
SecureID commands are described in detail in Part 2 of the AMI Configuration Commands Refer-
ence Manual.
8.1.1.2.5 Configuring SecurID on a Switch

The following sections describe how to install SecurID on a switch.
8.1.1.2.5.1 Installing the Server Software

A SecurID server must be run to implement the authentication method for users of the switch.
Refer to the Security Dynamics ACE/Server or Client user’s manuals for instructions about
installing the software on the server.
Security
8.1.1.2.5.2 Transferring the Configuration File
As described earlier, the server and the switches need to maintain some common configura-
tion parameters. The desired configuration information is specified in the sdconf.rec file
when the server is installed. Once you have installed the server software, copy this file to the
switch using the AMI command security securid getconf. This command uses either
the FTP protocol or the TFTP protocol to transfer the specified file from the specified server to
the switch.
This getconf command can be executed only

NOTE by users with admin privileges.

Security
This configuration file is read and information is stored in the FLASH so that they persist
across reboots. Additionally, the first time the switch contacts the server, it receives a node
secret file, which is a string of about 16 bytes. This string, which is known only to the server
and this switch, is used in encrypting messages between the server and the switch and is also
stored in the FLASH.
Even though this information is stored in the

NOTE FLASH, it cannot be accessed using any of the
system filesystem commands. The
security login securid getconf, delete,
and show commands must be used.
If the configuration file is not found, if the wrong file was copied, or if the file is corrupted,
then the SecurID service will not work. An appropriate error message is logged to the console,
and the user trying to log in is denied access.
8.1.1.2.5.3 Editing the Server Configuration File

Once SecurID is running, it is possible to make changes to the sdconf.rec file for the
server(s)/ clients. The file can be modified on the server itself by running sdsetup again and
changing the desired parameters.
After the changes are made on the server, use the

NOTE security securid getconf AMI command
to copy the updated sdconf.rec file back to all
of the switches that are using that file.
A sample of the contents of a sdconf.rec file and a description of the parameters follows:
Server License and Configuration
LICENSE CREATION: Mon July 27 15:06:08 1998

LICENSE ID: 96012648
ACE/Server VERSION: v 2.1.104
FILE OWNERSHIP: root
SETUID BIT: off
CLIENT RETRY: 5 times
CLIENT TIMEOUT: 5 sec
DES ENCRYPTION: allowed and enabled
DURESS MODE: not allowed
NetSP CLIENTS: not allowed

Security
EXTENDED TACACS: disabled

TACACS PLUS: disabled
MASTER SERVER: linus
MASTER SERVER ADDRESS: 204.95.89.107
SLAVE SERVER: allowed but not configured
AUTHENTICATION SERVICE: securid
PORT NUMBER: 1024
ADDRESSES: By name in host file or name service
BAD PASSCODES: 3
RESPONSE DELAY: 2
TOKENS IN LICENSE: 25
LICENSE CONFIGURATION
This license was created for:
Fore Systems, Inc.
1000 FORE Drive
Warrendale, PA 15086
These parameters are defined as follows:
Parameter Description
LICENSE CREATION The date the ACE/server license was created.
LICENSE ID The ACE/server license ID.
ACE/Server VERSION The version number of the server software.
FILE OWNERSHIP The account that owns the ACE/server files; i.e., that has permissions to run the ACE/
server administration programs. root is the default, but the login of any other adminis-
trator can be specified here.
SETUID BIT If set, allows anyone who can run the programs to run them with the permissions of the
files’ owner.
CLIENT RETRY Specifies how many times a client attempts to establish communications with the server
Security
before reporting an error.
CLIENT TIMEOUT Specifies how many seconds should elapse between attempts to establish client-server
communications.
DES ENCRYPTION Shows if DES or Security Dynamics encryption is used for client-server communications.
DURESS MODE Allows a user to enter a special PIN to signal that he or she is being forced to log in by
an unauthorized person.
NetSP CLIENTS Used for IBM’s Network Security Program.
EXTENDED TACACS Indicates if the server supports XTACACS clients.
TACACS PLUS Indicates if the server supports TACACS+ clients.
MASTER SERVER The name of the master server.

Security
MASTER SERVER ADDRESS The IP address of the master server.
SLAVE SERVER Indicates that the slave server may be installed. If it is installed, its name and IP address
become part of the configuration record.
AUTHENTICATION SERVICE The name of this authentication service as it appears in the /etc/services or in a NIS
Services file.
PORT NUMBER The UDP port number that has been assigned for the use of this service.
ADDRESSES Indicates how network devices shall have the IP addresses resolved.
BAD PASSCODES The number of failed login attempts with incorrect passcodes, after which the server
puts the token associated with that userid into Next Token Mode. When the token is in
this mode, the next time the same user tries to log in and gets the passcode correct, he or
she is also prompted to enter the next token code; i.e., the one that appears next on the
token after 60 seconds.
RESPONSE DELAY Indicates how long an authentication request should be held (by the server) before a
response is returned to the client.
TOKENS IN LICENSE The number of tokens for this license.
LICENSE CONFIGURATION Information about the owner of the license.
8.1.1.2.5.4 An Example Login Using SecurID

Once the SecurID server is set up and the configuration file has been copied to the switch, all
users who are configured for securID will see a password prompt after entering their user-
name when they attempt to log in to that switch. In the following example, a userid eng has
been created on a switch that has the sdconf.rec file in FLASH and that has the authentica-
tion method for userid eng set to securid. The following is a transcript of how the user logs
in, with the switch output in plain courier font and the user input in bold courier font:
localhost% telnet fishtank

Trying 172.21.5.103...
Connected to switchtestlf3.
Escape character is '^]'.
S_ForeThought_6.1.0 FCS (1.49367) (asx4000) (fishtank)

login: eng <ENTER>
Enter PASSWORD: <PIN><Code on SecurID Token> <ENTER>
Password:
If the passcode entered by the user is accepted, a session to the switch is started. If the server
denies access to the user, the following message appears on the screen: Login incorrect.
After three unsuccessful attempts, the telnet connection is torn down by the switch.

Security
8.1.1.3 Kerberos Authentication

ForeThought 6.0.x or greater supports Kerberos V5 authentication. Kerberos is a secret-key net-
work authentication protocol designed to authenticate requests for network resources. Kerberos
was created and developed at Massachusetts Institute of Technology (MIT) to provide a means
of verifying the identities of a workstation user or a network server, over an unsecure network.
Kerberos performs authentication as a trusted third party service (Kerberos server) using the
Data Encryption Standard (DES) for encryption. This data encryption method protects com-
munications between the server and the switch. When a user sends a request to the Kerberos
server for access to a switch, the server issues a ticket to the user that links that user to the
switch. This “ticket” is a record that contains the user’s identification and a temporary encryp-
tion key (session key). The session key is necessary to successfully use that ticket in an authen-
tication exchange. The user receives this ticket and transmits it back to the server with the
current time stamped. The session key, now shared by the user and server, is used to authenti-
cate the user to the switch. The ticket and the ticket’s session key is known as a credential.
Kerberos allows a user to log in to a switch via a DES encrypted telnet session with Kerberos
authenticated session key exchange.
Because Kerberos does not protect SNMP, ILMI,

NOTE remote AMI, or Element Manager access to the
switch, it is recommended that users either
employ IP filtering as a selective mechanism to
allow SNMP changes, or disable SETs from the
network entirely.
8.1.1.3.1 Kerberos Server

The Kerberos server can run on a UNIX system or on a Windows NT system. Each FORE
switch has a defined set of authorized users. See Part 2 of the AMI Configuration Commands
Reference Manual for more information. If a user requests a DES encrypted telnet session to a
switch, that user is not permitted access to that switch until his or her identity is validated.
Security
A Kerberos server maintains a database of user, server, and password information. When a
user attempts to access a switch via a DES encrypted telnet session, the switch configured for
Kerberos authentication verifies the session key and encrypts and decrypts messages sent
between the switch and the telnet session.
8.1.1.3.2 Kerberos AMI Commands

The security kerberos AMI command allows Kerberos configuration on the switch. The
Kerberos commands are described in detail in Part 2 of the AMI Configuration Commands Refer-
ence Manual.

Security
8.1.1.3.2.1 Configuring Kerberos Authentication on a Switch

The following tasks should be completed to configure Kerberos authentication:
• Install the Kerberos server software (Section 8.1.1.3.2.2)
• Configure the Kerberos server database (Section 8.1.1.3.2.3)
• Transfer the secret key file (Section 8.1.1.3.2.4)
• Set the time on the switch (Section 8.1.1.3.2.5)
For detailed information on the AMI commands

NOTE for configuring Kerberos authentication, see Part
2 of the AMI Configuration Commands Reference
Manual.
8.1.1.3.2.2 Installing the Kerberos Server Software

A Kerberos database server must be set up to implement the Kerberos authentication method
for users of the switch. A Kerberos server and database program running on a network host is
often referred to as the KDC (Key Distribution Center). Refer to the documentation that came
with your Kerberos software for instructions about installing the software on the server.
8.1.1.3.2.3 Configuring the Kerberos Server Database

All users and switches requiring Kerberos authentication must be registered in the Kerberos
server database. This allows hosts and the Kerberos server in your realm to communicate and
authenticate to each other. Kerberos uses the IP address and qualified host name to build a prin-
cipal name. A principal is a string consisting of the name or type of service being offered, and the
name of the host offering this service.
A key table (keytab) database (v5srvtab file) is generated for the principals entered into the
Kerberos server. This file contains service keys known only by the KDC and the switch.
8.1.1.3.2.4 Transferring the Secret Key File

Once the v5srvtab file has been generated with all the secret keys for the switch and its inter-
faces, the file should be installed on the switch using FTP. Use the security kerberos
v5srvtab install AMI command to transfer the specified file from the server to the switch.
security kerberos v5srvtab install ftp://164.144.22.2/kerberos/

v5srvtab

Security
This install command can be executed only

NOTE by users with admin privileges.
This key file is read only. Information is stored in the FLASH so that Kerberos configuration
persists across reboots.
Even though this information is stored in the

NOTE FLASH, it cannot be accessed using any of the
system filesystem commands. The
security kerberos show command must be
used.
Authentication fails if the key file is not found, if the wrong file was installed, or if the file is
corrupt. In these cases, an appropriate message is logged to the console, and access is denied.
8.1.1.3.2.5 Setting the Correct Time

For Kerberos authentication to work, it is important to set the correct time on the switch using
the system date AMI command. Kerberos does not utilize the time zone, but uses the
switch’s clock directory which is in GMT. The system date command uses the time zone to
convert the local time zone to GMT. The time set on the switch must be within five minutes of
the time on the KDC (or to whatever time change defined on the KDC). Otherwise, authentica-
tion will fail since Kerberos will assume that someone is trying a replay attack. See Part 3 of
the AMI Configuration Commands Reference Manual for more information.
8.1.1.3.2.6 Adding IP Address-to-Hostname Mappings

Once Kerberos is running, IP address-to-hostname mappings (for the local machine’s inter-
face) can be added or deleted from the CDB by entering the security kerberos host
new or security kerberos host delete AMI commands. The maximum number of
Security
entries allowed in the host table is 10.See Part 2 of the AMI Configuration Commands Reference
Manual.
security kerberos host new 169.144.44.4 guppy
security kerberos host new 169.144.33.3 shark

Security
Enter the following to display the IP address-to-Hostname mapping you just added:
security kerberos host show
IP Address Hostname
169.144.44.4 guppy
169.144.33.3 shark
8.1.1.3.2.7 Establish an Encrypted Telnet Session

A request for a DES encrypted telnet session allows users to be authenticated by their
Kerberos credentials before a Telnet session with a switch is established. Once the Kerberos
server is set up and the secret key file has been installed on the switch, all users who are con-
figured for Kerberos authentication will see a message when attempting to telnet into that
switch.
8.1.1.3.3 An Example Login Using Kerberos Authentication

Once the Kerberos server is set up and the key file has been installed on the switch, users who
are set up for Kerberos authentication will see a message when attempting to telnet into that
switch. In the following example, a userid jsmith has been created on a switch that has the
v5srvtab file in FLASH. The following is a transcript of how the user logs in, with the switch
output in plain courier font and the user input in bold courier font:
telnet -x fishtank
Trying 172.19.4.20...
Connected to fishtank.
[ Kerberos V5 accepts you as ``[email protected]'' ]
S_ForeThought_6.0.0 Development (1.29680+30d) (asx1000) (fishtank)
If Kerberos authentication fails, the following message displays:
Trying 172.19.4.20...
Connected to fishtank.
Authentication negotiation has failed, which is required for encryption. Goodbye.
After Kerberos authentication, you are still prompted for your AMI userid unless a default
profile is supplied for Kerberos. See Section 8.1 for more information on userids and default
profiles.

Security
8.1.2 AMI Command Privileges

The network administrator can configure two different levels of AMI command privileges for
userids: admin privileges, user privileges, and readonly privileges. Each method is described
in the following sections.
8.1.2.1 Admin Privileges

A person whose userid is configured with admin privileges is allowed to access and use all of
the AMI commands. This level of privileges should be reserved for the system administrator.
8.1.2.2 User Privileges

A person whose userid is configured with user privileges is allowed to access and use all
AMI commands, except the following:
• all debug commands
• all security commands
• hardware dualscp synchronize
• hardware netmod application upgrade
• system batch show
• system batch create
• system batch run
• system cdb init
• system cdb reset
• system filesystem init
• system license
• system startup
Security
• system upgrade
8.1.2.3 Readonly Privileges

A person whose userid is configured with readonly privileges is allowed to access AMI
commands for read-only purposes only.

Security
8.1.3 AMI Access Levels

The network administrator can configure four different levels of AMI access for userids: con-
sole, telnet, http, and snmp. Each method is described as follows:
• Console Access: A person whose userid has a serial level of access is allowed to
login to the switch only via the serial port.
• Telnet Access: A person whose userid has a network level of access is allowed to
login to the switch only via telnet.
• HTTP Access: A person whose userid has an http level of access is allowed to
login to the switch only via the web.
• SNMP Access: A person whose userid has an snmp level of access is allowed to
login to the switch only via SNMP.
8.1.4 Userid Password

The network administrator may also assign a password to a userid using the security
login password command. When a user logged in with user privileges wants to modify
the password, he or she must correctly enter the old password first before typing the new
password. However, a user logged in with admin privileges can change any userid password
without first entering the old local password, except their own. The maximum size is for a
password is 512 characters. Any characters are allowed, except the colon (:) character.
8.1.5 Privilege Level for Unlisted Users

One of the login scenarios in Table 8.2 allows users who are not listed in the switch to login,
provided that the SecurID or Kerberos server is accessible and provided that the SecurID or
Kerberos server accepts the userid. The network administrator may assign a privilege level of
user (access to only a certain subset of the AMI commands) or a privilege level of admin
(access to all AMI commands) for these users via the security login defaults command.

Security
8.2 IP Filtering
The IP filtering feature lets the network administrator limit access to the control port of the
switch to prevent unauthorized access to the switch. The switch performs filtering on incom-
ing IP packets by determining if there is a match between the packet’s header source address
and this table of authorized incoming IP addresses. If the addresses match, the packets are
accepted, provided that they meet the requirements set up by the other IP filtering flags; oth-
erwise, they are rejected. Statistics are kept of the number of rejected IP packets and about the
last IP packet that was rejected.
8.2.1 Authorized IP Address Table

Using the security ipfiltering new command, the administrator can create an entry in a
table of authorized IP addresses from which IP packets will be accepted. When the adminis-
trator creates an IP address entry in the table, he or she can apply a mask to specify a wildcard
range of allowable addresses. For example, an IP address of 163.26.54.6 with a mask of
255.255.255.255 means only address 163.26.54.6 is accepted against that entry. However, an IP
address of 163.26.54.6 with a mask of 255.255.0.0 means addresses 163.26.*.* are accepted
against that entry. The table can contain a maximum of 32 entries.
The administrator can also delete entries from the table and display the contents of the table.
The table will persist across a reboot. (See Part 2 of the AMI Configuration Commands Reference
Manual for more information.)
CAUTION When the authorized IP address table is empty,

all addresses are accepted. This is the default
state. Therefore, it is recommended that at least
one address be entered into the table. Otherwise,
anyone may access the switch via the control
Security
port. The address you enter must be the address
of the machine you are using. Otherwise, you
will lock yourself out of the switch.
8.2.2 IP Filtering Flags

There are three IP filtering flags that can be configured to limit IP access in other ways: ssr,
lsr, and all. These flags are set to allow or disallow IP packets using the security
ipfiltering flags -ssr, security ipaccess flags -lsr, and security ipaccess
flags -all commands. (See Part 2 of the AMI Configuration Commands Reference Manual for
more information about these commands.)

Security
8.2.2.1 Strict Source Routing Flag

Strict source routing is a setting within each IP packet. When enabled, it means that the sender
is providing the specific path that the IP packet must use. Because source routing allows any-
one to send arbitrary packets anywhere, it may be desirable to reject these types of packets for
security reasons.
If the ssr flag is set to allow, all incoming IP packets that are strict source routed are
accepted, provided that they match an IP address in the table of authorized addresses. If the
ssr flag is set to disallow, all incoming IP packets that are strict source routed are rejected,
even if they match an IP address in the table of authorized addresses.
8.2.2.2 Loose Source Routing Flag

Loose source routing is a setting within each IP packet. When enabled, it means that the
sender is providing a specific list of IP addresses that the packet must pass through, but the
packet may also pass through other routers between any two addresses in the list. Because
source routing allows anyone to send arbitrary packets anywhere, it may be desirable to reject
these types of packets for security reasons.
If the lsr flag is set to allow, all incoming IP packets that are loose source routed are
accepted, provided that they match an IP address in the table of authorized addresses. If the
lsr flag is set to disallow, all incoming IP packets that are loose source routed are rejected,
even if they match an IP address in the table of authorized addresses.
8.2.2.3 All Flag

If the all flag is set to allow, all incoming IP packets are accepted, provided that they match
an IP address in the table of authorized addresses. Allowing all is the default setting. If the
all flag is set to disallow, all incoming IP packets are rejected, even if they match an IP
address in the table of authorized addresses.
8.2.3 IP Access Statistics

The network administrator can display the total number of IP packets that have been filtered
since the switch was rebooted using the security ipfiltering statistics command.
This command also displays information about the last IP packet that was dropped. This
information includes the following:
• the reason that the last IP packet was dropped
• the system time at which the last IP packet was dropped
• the name of the interface on which the last dropped IP packet was received
• the IP address contained in the source field of the header of the last IP packet that
was dropped

Security
8.3 NSAP Filtering

NSAP filtering provides a mechanism for filtering calls based on a combination of the calling
(source) and called (destination) addresses, as well as the incoming and outgoing UNIs.
Each UNI may have one address filter for incoming call setups, and one for outgoing call set-
ups. If a call setup is routed from incoming UNI A to outgoing UNI B, A’s incoming call filter
and B’s outgoing call filter are applied. A call setup message must be accepted by both filters if
both are present.
SPVCs, SPVPs, PVPs, and PVCs are not

NOTE supported by NSAP filtering.
This feature also has a filter lookup mechanism which provides statistics about calls that were
filtered and information about the last call that was rejected by the filtering process.
8.3.1 Filters and Templates

Each filter is composed of an ordered set of templates. A template consists of the following:
• an incoming UNI (port/VPI)
• a source NSAP address/mask
• an outgoing UNI (port/VPI)
• a destination NSAP/Mask
• an action to either accept or reject the UNI or NSAP address
Masks may be wildcarded using a 0. Ports, VPIs, and NSAP addresses may be wildcarded
Security
using an asterisk (*). If * is specified for the port, then any port is accepted by the filter. If * is
specified for the VPI, then any VPI is accepted by the filter. If * is specified for the NSAP
address, or if 0 is specified for the mask, then any NSAP address is accepted by the filter.
Templates within a filter are applied in the order in which they appear in the filter, not by
maximum prefix match. See Part 2 of the AMI Configuration Commands Reference Manual for
more information about how to create, modify, and delete filters and templates.

Security
8.3.2 NSAP Filtering Lookup

A filter lookup mechanism is provided which allows you to enter components of a call setup
message to test whether a call setup attempt with the supplied addresses and ports would be
accepted or rejected by a specific filter. The switch returns an answer of accepted or
rejected and the index number of the template that accepted or rejected the information
entered. If the information does not match any of the existing templates, then an answer of
rejected and address unknown is given. When this feature is used, none of the statistics
are incremented and no traps are sent.
This command allows the user to enter components of a call setup message to test whether a
call setup attempt with the supplied addresses and ports would be accepted or rejected by a
specific filter.
myswitch:security nsapfiltering filters-> lookup

Usage:
[-filterid] <FilterId> Filter ID/Name
[[-srcatmif] <AtmIf>] Src AtmIf (default: ANY)
[[-srcvpi] (unsigned|ANY)] Src VPI (default: ANY)
[-srcnsap] <NSAP Address> Src NSAP
[[-dstatmif] <AtmIf>] Dst AtmIf (default: ANY)
[[-dstvpi] (unsigned|ANY)] Dst VPI (default: ANY)
[-dstnsap] <NSAP Address> Dst NSAP
usage parameters.
The switch returns an answer of either accepted or rejected, and the template-id of the
specific template that accepted or rejected the tested call setup message. If the message does
not match any of the templates in the filter, the switch returns an answer of rejected and
address unknown.
This command is simply a test. Therefore, if a

NOTE message is rejected, it does not increment the
address validation counters, enter the result in
the table that shows the last failure, nor send a
trap.

Security
8.3.3 NSAP Filtering Statistics

To display statistics for NSAP filtering, use the security nsapfiltering interfaces
callstatistics command:
myswitch:security nsapfiltering interfaces-> callstatistics

AtmIf VPI Direction Accepts Rejects Unmatched
----- --- --------- ------- ------- ---------
2A1 0 incoming 0 0 0
2A1 0 outgoing 0 0 0
2B1 0 incoming 0 0 0
2B1 0 outgoing 0 0 0
5CTL 0 incoming 0 0 0
5CTL 0 outgoing 0 0 0
See Part 2 of the AMI Configuration Commands Reference Manual for field descriptions.
NSAP filtering statistics and traps include:
Security
• counts of accepted calls
• calls rejected because of an explicit match with a reject template
• calls rejected because of no match with any template
A trap is sent when rejected calls occur more frequently than the user-specified limit.

Security

This chapter describes how to set up timing on a FORE ATM Switch. Topics covered include:
• Section 9.1 - Overview
• Section 9.2 - Timing Modes
• Section 9.3 - Switchclock
• Section 9.4 - Port Level Timing
• Section 9.5 - Timing Configuration Examples
9.1 Overview
When running voice or video over an ATM network, the equipment on the network should be
timed so that all of the equipment is transmitting and receiving data in a synchronized man-
ner. Even though ATM is asynchronous, this only refers to the cell-level transmission that
occurs in the ATM Layer. The physical layer underlying the cells is synchronous. Therefore, a
time reference signal must be distributed to every element in the network to establish one
cohesive entity. This time reference on FORE ATM Switches is called a switchclock.
9.2 Timing Modes

There are two timing modes available on FORE ATM Switches: TCM and switch. TCM mode
means that all network modules that support distributed timing import their clock source
from the Timing Control Module (TCM).
You must have a TCM installed in your switch in

NOTE order to use this timing mode. See the CEC-Plus
User’s Manual for information about configuring
timing with a TCM. All examples in this chapter
Configuring Timing
assume that you do not have a TCM installed.
Switch mode means that all network modules that support distributed timing import their
clock source from the designated switchclock port. This is the default mode for any switch
that does not have a TCM installed. See Section 9.3 for more information about this mode.

Configuring Timing
9.3 Switchclock
If a TCM is not installed in the switch, all of the ports within a fabric use the switchclock as
their timing reference. The switchclock can be any port that is able to recover a clock (i.e., the
network module supports distributed timing). (To see if the network module supports distrib-
uted timing, use the command hardware netmod show and look for yes under the Timing
field.) In an ASX-1000, ASX-1200, or a TNX-1100, the port may be from another fabric in the
same chassis.
When the switchclock is set to any of the ports within the same fabric, that port’s clock is
exported to all of the network modules within the same fabric. On an ASX-1000, ASX-1200, or
a TNX-1100 only, the switchclock is also exported to the network modules in all of the other
fabrics in this chassis. You can then configure each of the other fabrics to use that clock. In this
way, all of the network modules in all of the fabrics will use the same timing source.
9.3.1 Failover of the Switchclock

You can configure a primary switchclock and a secondary switchclock. If the primary clock is
valid, it is used. If the primary switchclock fails, the secondary clock is used. When the pri-
mary clock returns, it is used again as the clock source.
If both the primary and secondary values fail, the switch fabric uses the crystal of the first
available timing network module as the switchclock, going from A to D. For example, if net-
work module A supports distributed timing, then the crystal from A is used as the switch-
clock. As another example, if network module A is not installed, and B and C do not support
distributed timing, but D does support distributed timing, then the crystal from D is used as
the switchclock. If the crystal of the network module has to be used for timing support, other
ports from other fabrics can use that particular network module crystal as their timing refer-
ence. The network module crystal is of Stratum 4 level accuracy.
Failover of the switchclock happens automatically when any of the following events occur:
• A carrier detect or carrier loss occurs on any port in this switch fabric.
• A network module is swapped in or out.
• A switch fabric is swapped in or out.

Configuring Timing
9.4 Port Level Timing

In addition to the switchclock, there is also a configurable parameter for each port called a
transmit clock (txclock). Each port’s txclock can be configured to use either the network clock or
the internal clock.
• When the txclock is set to network, the clock that is recovered from the receive
line of a port is used to drive the transmit line of that port.
• When the txclock is set to internal, the internal clock, or switchclock, is used to
drive the transmit line of that port.
The Timing field under hardware port sonet

NOTE show displays N/A for all series of OC-12
network modules because they always use
internal timing.
Configuring Timing

Configuring Timing
9.5 Timing Configuration Examples

This section provides examples of how to set the switchclock on various types of switches. For
more information about the specific timing commands, see Part 3 of the AMI Configuration
Commands Reference Manual.
If you have a TCM installed in your switch, see the CEC-Plus User’s Manual for information
about configuring timing with a TCM. All examples in this chapter assume that you do not
have a TCM installed.
9.5.1 Configuring Timing on an ASX-200BX, TNX-210, or

ASX-200WG
This example assumes that you are going to use 1A1 as your primary clock and 1B1 as your
secondary clock. Use the following ATM Management Interface (AMI) commands to set your
primary and secondary switchclocks:
system timing modify -primaryclock 1a1
system timing modify -secondaryclock 1b1

(Single Timing Domain)
secondary clock. The only limitation on an ASX-1000, ASX-1200, or a TNX-1100 is that you
must configure both the primary and secondary clocks to be on the same fabric. On all fabrics
installed in the switch, use the following AMI commands to set your primary and secondary
clocks:

Configuring Timing

(Multiple Timing Domains)
This example assumes that you are going to use 1C1 as your primary clock and 1D1 as your sec-
ondary clock for fabrics 1 and 2 and that you are going to use 3A1 as your primary clock and 3B1
as your secondary clock for fabrics 3 and 4. The only limitation on an ASX-1000, ASX-1200, or a
TNX-1100 is that you must configure both the primary and secondary export clocks to be on the
same fabric.
1. On fabrics 1 and 2, use the following AMI commands to set your primary and sec-
ondary clocks:
system timing modify -primaryclock 1c1
system timing modify -secondaryclock 1d1
2. On fabrics 3 and 4, use the following AMI commands to set your primary and sec-
ondary clocks:
9.5.4 Configuring Timing on a ForeRunnerLE 155 or LE 25

The ports in interface groups A, B, C, and the Port Expansion Module (PEM) (interface D) use
the crystal on the fabric as their timing source by default. Only the PEM can recover a clock
from any of its ports and export it to the other interfaces. If the PEM is installed, the primary
and secondary clocks be recovered from one of the PEM ports.
This example assumes that you are going to use 1D1 as your primary clock and 1D2 as your
secondary clock. Use the following AMI commands to set your primary and secondary clocks:
system timing modify -primaryclock 1d1
system timing modify -secondaryclock 1d2

Configuring Timing

Configuring Timing
9.5.5 Configuring Timing on an ASX-4000

secondary clock. Use the following AMI commands to set your primary and secondary
switchclocks:
On a channelized OC-12 port card in an

NOTE ASX-4000, distributed timing is recovered from
the OC-12 SONET port, not the DSX3 interface.
9.5.6 Configuring Timing on an ESX-3000

The ports in the interface groups on the ESX-3000 ATM port cards use their own crystal as
their timing source by default. Only the first port on each port card (e.g., 3C1, 5A1, etc.) can
recover a clock from any of its ports and export it to the other interface groups. The primary or
the secondary clock can be recovered from the first port on any port card. Only one timing
source can be configured per port card.
This example assumes that you are going to use 3B1 as your primary clock and 1C1 as your
secondary clock. Use the following AMI commands to set your primary and secondary clocks:
system timing modify -primaryclock 3b1
system timing modify -secondaryclock 1c1
In the example above, the primary timing source is set to 3B1. The switch exports the timing
source from the ATM port card in slot 3 to the entire chassis.

Configuring SNMP
The switch control software for the FORE ATM switches includes an SNMP agent. The SNMP
agent enables the remote monitoring and configuration of these switches.
A.1 SNMP Indexing

For a description of the SNMP indexing scheme used on FORE switches, please see the ATM
Switch Diagnostics and Troubleshooting Manual.
ATM Switch Network Configuration Manual A-1

Configuring SNMP
A.2 SNMP Traps

SNMP traps are used to update the state of the network automatically to remote network
management hosts. The SNMP agent on the switch supports several SNMP traps.
The traps generated by the switch’s SNMP agent can be sent to as many destinations as
needed. These destinations are configurable via the ATM Management Interface (AMI). Each
destination must be an IP address of a network management host. The network management
host specified for a trap destination can be any host with which the switch has connectivity.
This means that the host does not have to be a directly connected ATM host. It can be on any
attached network. Table A.1 describes the supported traps.
Table A.1 - SNMP Traps Supported on the FORE Switches
Trap
Trap Name Description
Number
0 asxSwLinkDown An asxSwLinkDown trap signifies that the sending
protocol entity recognizes a failure in one of the ATM
Switch links that is connected to another switch.
1 asxSwLinkUp An asxSwLinkUp trap signifies that the sending proto-
col entity recognizes that one of the ATM Switch links
that is connected to another switch has come up.
2 asxHostLinkDown An asxHostLinkDown trap signifies that the sending
protocol entity recognizes a failure in one of ATM
Switch links that is connected to a host.
3 asxHostLinkUp An asxHostLinkUp trap signifies that the sending pro-
tocol entity recognizes that one of the ATM Switch
links that is connected to a host has come up.
4 asxNetModuleDown An asxNetModuleDown trap signifies that the sending
Switch network modules, that is identified by the
board and the module numbers. This is probably
caused by a hot-swap of a network module.
5 asxNetModuleUp An asxNetModuleUp trap signifies that the sending
protocol entity recognizes a new operational ATM
Switch network module, that is identified by the board
and the module numbers. This is probably caused by a
hot-swap of a network module.
A-2 ATM Switch Network Configuration Manual

Configuring SNMP
Table A.1 - SNMP Traps Supported on the FORE Switches (Continued)
Configuring SNMP
Trap
Number
6 asxPsInputDown This trap alerts that one ATM switch power supply
failed due to failure in the input voltage. The power
supply that failed is identified by the power supply
index. Note that an input voltage may be out of specifi-
cation and may not cause a power supply failure if
high loads are not applied.
7 asxPsInputUp This trap alerts that one ATM switch power supply
that had an input failure is up. The power supply that
is back up is identified by the power supply index.
9 asxPsOutputDown This trap alerts that one ATM switch power supply
output or the power supply was physically removed.
The power supply that failed is identified by the
power supply index.
10 asxPsOutputUp This trap alerts that one ATM switch power supply
that had an output failure or was removed is now up.
The power supply that is back up is identified by the
power supply index.
22 asxFanBankDown This trap alerts that one ATM switch fan bank failed.
The fan bank that failed is identified by the fan bank
index.
23 asxFanBankUp This trap alerts that one ATM switch fan bank is up.
The fan bank that is back up is identified by the fan
bank index.
28 asxLinkDown This trap alerts that the link that is identified by
{hwPortBoard, hwPortModule, hwPortNumber} was
configured up but lost its carrier (or the framing bit)
and is currently down.
29 asxLinkUp This trap alerts that the link that is identified by
{hwPortBoard, hwPortModule, hwPortNumber} is
back up.
30 asxSpansDown This trap alerts that the SPANS signalling on the link
that is identified by the sigPathPort and sigPathVPI
failed.

Configuring SNMP
Trap
Number
31 asxSpansUp This trap alerts that the SPANS signalling on the link
that is identified by the sigPathPort and sigPathVPI is
up.
32 asxTempSensorOverTemp This trap alerts that one of the temperature sensors
indicates over temperature. The temperature sensor is
identified by the temperature sensor index.
33 asxTempSensorRegularTemp This trap alerts that one of the temperature sensors
indicates regular temperature. The temperature sensor
is identified by the temperature sensor index.
34 asxFabricTemperature This trap alerts that one of the temperature sensors
OverTemp indicates over temperature. The temperature sensor is
identified by the temperature sensor index.
35 asxFabricTemperature This trap alerts that one of the temperature sensors
RegularTemp indicates regular temperature. The temperature sensor
is identified by the temperature sensor index.
36 asxSonetLOSDetected This trap indicates that the specified SONET port is
experiencing Loss Of Signal. Bellcore Document
TA-NWT-000253 Section 6.3.1.1.1 states that, “A
SONET NE shall declare a LOS failure when the LOS
defect persists for 2.5 (± .5) seconds, or when a LOS
defect is present and the criteria for LOF failure decla-
ration have been met.”
37 asxSonetLOSCleared This trap indicates that the LOS condition identified by
trap asxSonetLOSon has been cleared.
38 asxSonetPathLabelOn This trap indicates that the specified SONET port is
receiving and errored C2 Path Label byte. Reference
Bellcore Document TA-NWT-000253 Section 3.3.2.3
and 6.3.1.1.8 the Path Label (C2) byte should have the
value 0x13.
39 asxSonetPathLabelOff This trap indicates that the Errored Path Label (C2)
byte error condition signalled by the asxSonetPath-
LabelOn trap has been cleared.

Configuring SNMP
Configuring SNMP
Trap
Number
40 asxSonetLineAISDetected This trap indicates that the specified SONET port is
receiving a Line level Alarm Indication Signal from the
far-end equipment.
41 asxSonetLineAISCleared This trap indicates that the Line AIS error condition
signalled by the asxSonetLineAISon trap has been
cleared.
46 asxDS3PLCPYellowDetected This trap indicates that the specified DS3 port has
detected incoming Yellow Alarm.
47 asxDS3PLCPYellowCleared This trap indicates that the specified DS3 port has
detected clearance of incoming Yellow Alarm.
48 asxDS3PLCPLOFDetected This trap indicates that the specified DS3 port has
detected incoming LOF Alarm.
49 asxDS3PLCPLOFCleared This trap indicates that the specified DS3 port has
detected clearance of incoming LOF Alarm.
50 asxDS3LOFDetected This trap indicates that Loss Of Frame(LOF) is
detected on the incoming signal.
51 asxDS3LOFCleared This trap indicates that Loss Of Frame is cleared on the
incoming signal.
52 asxDS3AISDetected This trap indicates that AIS Alarm is detected on the
incoming signal.
53 asxDS3AISCleared This trap indicates that AIS Alarm is cleared on the
incoming signal.
60 asxDS1PLCPYellowDetected This trap indicates that the specified DS1 port has
detected an incoming Yellow Alarm.
61 asxDS1PLCPYellowCleared This trap indicates that the specified DS1 port has
detected clearance of an incoming Yellow Alarm.
62 asxDS1PLCPLOFDetected This trap indicates that the specified DS1 port has
detected an incoming LOF Alarm.
63 asxDS1PLCPLOFCleared This trap indicates that the specified DS1 port has
detected clearance of an incoming LOF Alarm.
64 asxDS1YellowDetected This trap indicates that Yellow Alarm is detected on
the incoming signal.

Configuring SNMP
Trap
Number
65 asxDS1YellowCleared This trap indicates that Yellow Alarm is cleared on the
incoming signal.
66 asxDS1AISDetected This trap indicates that AIS Alarm is detected on the
incoming signal.
67 asxDS1AISCleared This trap indicates that AIS Alarm is cleared on the
incoming signal.
68 asxDS1LOSDetected This trap indicates that LOS Alarm is detected on the
incoming signal.
69 asxDS1LOSCleared This trap indicates that LOS Alarm is cleared on the
incoming signal.
70 asxDS1LOFDetected This trap indicates that LOF Alarm is detected on the
incoming signal.
71 asxDS1LOFCleared This trap indicates that LOF Alarm is cleared on the
incoming signal.
74 asxDS3FERFDetected This trap indicates that FERF Alarm is detected on the
incoming signal.
75 asxDS3FERFCleared This trap indicates that FERF Alarm is cleared on the
incoming signal.
78 asxE3YellowDetected This trap indicates that the Yellow Alarm is being
79 asxE3YellowCleared This trap indicates that Yellow alarm has cleared on the
incoming signal.
80 asxE3OOFDetected This trap indicates that Out Of Frame (OOF) is
81 asxE3OOFCleared This trap indicates that Loss Of Frame is cleared on the
incoming signal.
82 asxE3AtmLCDDetected This trap indicates that the specified E3 port is experi-
encing Loss of Cell Delineation (LCD). An LCD failure
is declared when the LCD defect persists for a period
of 2.5 +/- 0.5 seconds.

Configuring SNMP
Configuring SNMP
Trap
Number
83 asxE3AtmLCDCleared This trap indicates that the LCD failure identified by
trap asxE3AtmLCDDetected has been cleared. An LCD
failure is cleared when the LCD defect is absent for 10
+/- 0.5 seconds.
86 asxE3PLCPYellowDetected This trap indicates that the specified E3 port has
detected incoming PLCP Yellow Alarm.
87 asxE3PLCPYellowCleared This trap indicates that the specified E3 port has
detected clearance of incoming PLCP Yellow Alarm.
90 asxE1YellowDetected This trap indicates that the Yellow Alarm is being
91 asxE1YellowCleared This trap indicates that Yellow alarm has cleared on the
incoming signal.
92 asxE1LOFDetected This trap indicates that LOF is being detected on the
incoming signal.
93 asxE1LOFCleared This trap indicates that LOF is cleared on the incoming
signal.
96 asxE1PLCPYellowDetected This trap indicates that the specified E1 port has
detected incoming PLCP Yellow Alarm.
97 asxE1PLCPYellowCleared This trap indicates that the specified E1 port has
detected clearance of incoming PLCP Yellow Alarm.
98 asxE1PLCPLOFDetected This trap indicates that the specified E1 port has
detected incoming PLCP LOF Alarm.
99 asxE1PLCPLOFCleared This trap indicates that incoming PLCP LOF alarm has
been cleared on the specified E1 port.
100 asxE1LOSDetected This trap indicates that the specified E1 port has
detected incoming LOS Alarm.
101 asxE1LOSCleared This trap indicates that incoming LOS alarm has been
cleared on the specified E1 port.
102 asxE1AISDetected This trap indicates that the specified E1 port has
detected incoming AIS Alarm.
103 asxE1AISCleared This trap indicates that incoming AIS alarm has been

Configuring SNMP
Trap
Number
104 asxE3AISDetected This trap indicates that the specified E3 port has
detected incoming AIS Alarm.
105 asxE3AISCleared This trap indicates that incoming AIS alarm has been
106 asxE3LOSDetected This trap indicates that the specified E3 port has
107 asxE3LOSCleared This trap indicates that incoming LOS alarm has been
108 asxE3PLCPLOFDetected This trap indicates that the specified E3 port has
detected incoming PLCP LOF Alarm.
109 asxE3PLCPLOFCleared This trap indicates that incoming PLCP LOF alarm has
been cleared on the specified E3 port.
112 asxJ2YellowDetected This trap indicates that Yellow Alarm is detected on
the incoming signal.
113 asxJ2YellowCleared This trap indicates that Yellow Alarm is cleared on the
incoming signal.
114 asxJ2AISDetected This trap indicates that AIS Alarm is detected on the
incoming signal.
115 asxJ2AISCleared This trap indicates that AIS Alarm is cleared on the
incoming signal.
116 asxJ2LOSDetected This trap indicates that LOS Alarm is detected on the
incoming signal.
117 asxJ2LOSCleared This trap indicates that LOS Alarm is cleared on the
incoming signal.
118 asxJ2LOFDetected This trap indicates that LOF Alarm is detected on the
incoming signal.
119 asxJ2LOFCleared This trap indicates that LOF Alarm is cleared on the
incoming signal.
120 asxDS3LOSDetected This trap indicates that the specified DS3 port has

Configuring SNMP
Configuring SNMP
Trap
Number
121 asxDS3LOSCleared This trap indicates that the incoming LOS Alarm has
been cleared on the specified DS3 port.
130 asxSonetLOFDetected This trap indicates that the specified SONET port is
experiencing Loss Of Frame (LOF) failure. An LOF
failure is declared when the LOF defect persists for a
period of 2.5 +/- 0.5 seconds, except when an LOS
defect or failure is present.
131 asxSonetLOFCleared This trap indicates that the LOF failure identified by
trap asxSonetLOFDetected has been cleared. The LOF
failure is cleared when the LOS failure is declared, or
when the LOF defect is absent for 10 +/- 0.5 seconds.
132 asxSonetLineRDIDetected This trap indicates that the specified SONET port is
experiencing Line Remote Defect Indication (LRDI). A
Line RDI failure is declared when the incoming Line
RDI defects lasts for 2.5 +/- 0.5 seconds.
133 asxSonetLineRDICleared This trap indicates that the Line RDI failure identified
by trap asxSonetLineRDIDetected has been cleared.
The Line RDI failure is cleared when no Line RDI
defects are detected for 10 +/- 0.5 seconds.
134 asxSonetPathAISDetected This trap indicates that the specified SONET port is
experiencing Path Alarm Indication Signal (PAIS). A
Path AIS failure is declared when the Path AIS defect
persists for 2.5 +/- 0.5 seconds.
135 asxSonetPathAISCleared This trap indicates that the Path AIS failure identified
by trap asxSonetPathAISDetected has been cleared. A
PAIS failure is cleared when the PAIS defect is absent
for 10 +/- 0.5 seconds.
136 asxSonetPathLOPDetected This trap indicates that the specified SONET port is
experiencing Loss Of Pointer (LOP). A LOP failure is
declared when the LOP defect persists for a period of
2.5 +/- 0.5 seconds.
137 asxSonetPathLOPCleared This trap indicates that the LOP failure identified by
trap asxSonetLOPDetected has been cleared. A LOP
failure is cleared when the LOP defect is absent for 10
+/- 0.5 seconds.

Configuring SNMP
Trap
Number
138 asxSonetPathUNEQDetected This trap indicates that the specified SONET port is
experiencing unequipped (UNEQ). A UNEQ failure is
declared when the UNEQ defect persists for a period
of 2.5 +/- 0.5 seconds.
139 asxSonetPathUNEQCleared This trap indicates that the UNEQ failure identified by
trap asxSonetUNEQDetected has been cleared. A
UNEQ failure is cleared when the UNEQ defect is
absent for 10 +/- 0.5 seconds.
140 asxSonetPathRDIDetected This trap indicates that the specified SONET port is
experiencing Path Remote Defect Indication (PRDI). A
Path RDI failure is declared when the incoming Path
RDI defects lasts for 2.5 +/- 0.5 seconds.
141 asxSonetPathRDICleared This trap indicates that the Path RDI failure identified
by trap asxSonetPathRDIDetected has been cleared.
The Path RDI failure is cleared when no Path RDI
defects are detected for 10 +/- 0.5 seconds.
142 asxSonetAtmLCDDetected This trap indicates that the specified SONET port is
experiencing Loss of Cell Delineation (LCD). A LCD
failure is declared when the LCD defect persists for a
period of 2.5 +/- 0.5 seconds.
143 asxSonetAtmLCDCleared This trap indicates that the LCD failure identified by
trap asxSonetAtmLCDDetected has been cleared. A
LCD failure is cleared when the LCD defect is absent
144 asxSonetAtmLineBIP- This trap indicates that the specified SONET port is
Detected experiencing Bit Interleaved Parity (BIP) errors. A Line
BIP failure is declared when the Line BIP defect per-
sists for a period of 2.5 +/- 0.5 seconds.
145 asxSonetAtmLineBIPCleared This trap indicates that the Line BIP failure identified
by trap asxSonetAtmLineBIPDetected has been
cleared. A Line BIP failure is cleared when the Line BIP
defect is absent for 10 +/- 0.5 seconds.
160 asxDS3IdleDetected This trap indicates that an Idle Maintenance Signal
(IDLE) is detected on the incoming signal.
A - 10 ATM Switch Network Configuration Manual

Configuring SNMP
Configuring SNMP
Trap
Number
161 asxDS3IdleCleared This trap indicates that an Idle Maintenance Signal
(IDLE) is cleared on the incoming signal.
162 asxDS3AtmLCDDetected This trap indicates that the specified DS3 port is expe-
riencing Loss of Cell Delineation (LCD). An LCD fail-
ure is declared when the LCD defect persists for a
163 asxDS3AtmLCDCleared This trap indicates that the LCD failure identified by
trap asxDS3AtmLCDDetected has been cleared. An
164 asxDS3PbitPerrDetected This trap indicates that the specified DS3 port is expe-
riencing P-bit Parity errors. A P-bit Parity Error failure
is declared when the P-bit Parity Error persists for a
165 asxDS3PbitPerrCleared This trap indicates that the P-bit Parity Error failure
identified by trap asxDS3PbitPerrDetected has been
cleared. A P-bit Parity Error failure is cleared when the
P-bit Parity Error defect is absent for 10 +/- 0.5 sec-
onds.
176 asxDS1PRBSDetected This trap indicates that PRBS pattern is detected on the
incoming signal.
177 asxDS1PRBSCleared This trap indicates that PRBS pattern is cleared on the
incoming signal.
178 asxDS1AtmLCDDetected This trap indicates that the specified DS1 port is expe-
riencing Loss of Cell Delineation (LCD). An LCD fail-
ure is declared when the LCD defect persists for a
179 asxDS1AtmLCDCleared This trap indicates that the LCD failure identified by
trap asxDS1AtmLCDDetected has been cleared. An
ATM Switch Network Configuration Manual A - 11

Configuring SNMP
Trap
Number
180 asxDS1CRCErrDetected This trap indicates that the specified DS1 port is expe-
riencing excessive CRC-6 errors. A CRC-6 Error failure
is declared when the CRC-6 Error persists for a period
of 2.5 +/- 0.5 seconds.
181 asxDS1CRCErrCleared This trap indicates that the excessive CRC-6 Error fail-
ure identified by trap asxDS1CRCErrDetected has
been cleared. A CRC-6 Parity Error failure is cleared
when the CRC-6 Error defect is absent for 10 +/- 0.5
seconds.
192 asxE3TrailChangeDetected This trap indicates that a change in the trail trace mes-
sage was detected on the incoming signal.
208 asxE1AtmLCDDetected This trap indicates that the specified E1 port is experi-
of 2.5 +/- 0.5 seconds.
209 asxE1AtmLCDCleared This trap indicates that the LCD failure identified by
trap asxE1AtmLCDDetected has been cleared. An LCD
+/- 0.5 seconds.
224 asxJ2RLOCDetected This trap indicates that Receive Loss of Clock (RLOC)
is detected on the incoming signal.
225 asxJ2RLOCCleared This trap indicates that Receive Loss of Clock (RLOC)
is cleared on the incoming signal.
226 asxJ2HBERDetected This trap indicates that High Bit Error Rate (HBER) is
227 asxJ2HBERCleared This trap indicates that High Bit Error Rate (HBER) is
cleared on the incoming signal.
228 asxJ2PAISDetected This trap indicates that Payload Alarm Indication Sig-
nal (PAIS) is detected on the incoming signal.
229 asxJ2PAISCleared This trap indicates that Payload Alarm Indication Sig-
nal (PAIS) is cleared on the incoming signal.

Configuring SNMP
Configuring SNMP
Trap
Number
230 asxJ2AtmLCDDetected This trap indicates that the specified J2 port is experi-
of 2.5 +/- 0.5 seconds.
231 asxJ2AtmLCDCleared This trap indicates that the LCD failure identified by
trap asxJ2AtmLCDDetected has been cleared. An LCD
+/- 0.5 seconds.
232 asxJ2TLOCDetected This trap indicates that Transmit Loss of Clock (TLOC)
is detected.
233 asxJ2TLOCCleared This trap indicates that Transmit Loss of Clock (TLOC)
is cleared.
250 asxTP25LOSDetected This trap indicates that LOS Alarm is detected on the
incoming signal.
251 asxTP25LOSCleared This trap indicates that LOS Alarm is cleared on the
incoming signal.
1024 asxOutputQueueCongested This trap indicates that the output queue for the given
priority has exceeded its dedicated length, and has
begun overflowing into the shared buffer space on the
network module.
1025 asxOutputQueueCellLoss This trap indicates that the output queue for the given
priority has overflowed and cells have been dropped.
1026 asxExtendedModeViolation This trap indicates that a series A or B network mod-
ule was inserted into a switch board running in
extended mode.
1027 asxNonextendedMode- This trap indicates that a series C or greater network
Warning module was inserted into a switch board running in
non-extended mode.
1028 q2931AVRejectTrap This trap is generated whenever any UNI3.x with
AddressValidation enabled rejects a Setup Request call
more than q2931AVRejectTrapThreshold times in any
given q2931AVRejectTrapPeriod.

Configuring SNMP
Trap
Number
1029 crConfMemoryOflow This trap is generated when the allocated call record
memory (as indicated by crMemoryAllocated) is
exceeded.
1030 crXfrPrimaryXfrFailed This trap is generated when the call record transfer to
the primary host (as indicated by crXfrPrimaryUrl)
fails.
1031 crXfrSecondaryXfrFailed This trap is generated when the call record transfer to
the secondary host (as indicated by crXfrSecond-
aryUrl) fails.
1032 crConfMemAllocFail This trap is generated when Callrecord functionality is
unable to allocate memory as specified by crMemory-
Allocated. This can happen when the crConfAdmin-
Status changes state from “off” or when the switch
reboots when Callrecords is configured “on”.
1033 crGeneralFailure This trap is generated when any of the callrecord
related functionality fails for any reason. One example
would be when the Callrecord Module fails to sched-
ule an interval timer.
1034 asxDualScpSyncFailure This trap indicates that automatic CDB synchroniza-
tion is disabled due to failures.
1035 asxDualScpSwitchOver This trap indicates that the backup SCP has taken con-
trol of the switch.
1036 asxDualScpHotSwap This trap indicates that an SCP hotswap insertion or
removal has occurred.
1037 asxVPAISDetected This trap indicates that the Alarm Indication Signal
(AIS) is detected on the incoming (terminating) virtual
path. This trap is generated once when the virtual path
is declared to be in the active AIS state.
1038 asxVPAISCleared This trap indicates that the Alarm Indication Signal
(AIS) has been removed from the incoming (terminat-
ing) virtual path. This trap is generated once when the
virtual path is declared to be in the inactive AIS state.

Configuring SNMP
Configuring SNMP
Trap
Number
1039 asxVPRDIDetected This trap indicates that the Remote Defect Indication
(RDI) is detected on the incoming (terminating) virtual
path. This trap is generated once when the virtual path
is declared to be in the active RDI state.
1040 asxVPRDICleared This trap indicates that the Remote Defect Indication
(RDI) has been removed from the incoming (terminat-
ing) virtual path. This trap is generated once when the
virtual path is declared to be in the inactive RDI state.
1041 asxNonextendedModeViola- This trap indicates that a Series D network module was
tion inserted into a switch board running in non-extended
mode. Multicast will not work on the Series D module
without removing all Series B modules and rebooting
the switch.
1042 asxUnsupportedNetwork- This trap indicates that a unsupported network mod-
Module ule was inserted into a switch.
1043 asxDualScpRedundancy This trap indicates that there is a change in the state of
dual SCP redundancy.
1049 asxIpFilterViolation This trap occurs when an incoming IP packet is unau-
thorized to enter the switch control port and has been
dropped.
1053 q2931AFRejectKnown This trap is generated whenever any q2931 UNI with
Address Filtering enabled rejects a Setup request
because the request matched a template with the
action “reject.” The variables sent in the trap identify
the source and destination UNI for the call.
1054 q2931AFRejectUnknown This trap is generated whenever any q2931 UNI with
Address Filtering enabled rejects a Setup request
because the address matched no template. The vari-
ables sent in the trap identify the source and destina-
tion UNI for the call.
1061 q2931CreationFailure This trap is generated whenever a switch fails to create
a UNI. This is most likely due to a resource limitation
on the switch.

Configuring SNMP
Trap
Number
1068 asxPsCurrentDown This trap alerts that one ATM switch power supply
had a current failure. The power supply that failed is
identified by the power supply index.
1069 asxPsCurrentUp This trap alerts that one ATM switch power supply
that had a current failure is now up. The power supply
that is back up is identified by the power supply index.
1070 asxPs5VoltDown This trap alerts that one ATM switch power supply
had a +5V failure. The power supply that failed is
identified by the power supply index.
1071 asxPs5VoltUp This trap alerts that one ATM switch power supply
that had a +5V failure is now up. The power supply
that is back up is identified by the power supply index.
1072 asxSwitchLoginDetected An asxSwitchLoginDetected trap signifies that a user
logged in on the switch.
1073 asxSwitchLoginFailed An asxSwitchLoginFailed trap signifies that a user’s
attempt to log in on the switch failed.
1074 pnniTdbGuardbandResrv- This trap is generated when the guardband memory
Fail reserve for any of the PNNI TDB (Topology Database)
related functionality like creation, modification, and
deletion on objects like node, PTSE, flags, internal pre-
fixes, external prefixes, etc. fails. The switch is low on
memory.
1075 pnniTdbInconsistentState This trap is generated when the PNNI TDB (Topology
Database) is in an unrecoverable error condition due to
MALLOC and other TDB related failures. When this
happens, the associated logical node is shut down and
this trap is sent. The switch has to be rebooted to bring
this PNNI logical node up again.
1077 asxShmem2OutputQueue This trap indicates that the output queue for the
Congested given priority has exceeded its dedicated length,
and has begun overflowing into the shared buffer
space on the network module.

Configuring SNMP
Configuring SNMP
Trap
Number
1078 asxShmem2OutputQueue This trap indicates that the output queue for the
CellLoss given priority has overflowed and cells have been
dropped.
1080 fabricLvl3Lookup This trap is generated when a fabric detects excessive
level 3 lookup errors. The level 3 lookup count is read
every second. If the lookup count is non-null for 5 con-
secutive seconds, then this trap is generated. If the
errors continue, this trap is generated continuously
every 5 seconds. This trap can be caused by: a miscon-
figured PVC; bad hardware.
1081 fabricCorrectedLookup This trap is generated when the SCP detects inconsis-
tent information in the HDCOMP lookup. The switch
has corrected the bad information in the HDCOMP
lookup ASIC’s data structures.
1090 spvcRerouteInitiated This trap is sent when a SPVC is being re-routed due to
a better path being found.
1091 asxQ2931Down This trap is generated whenever UNI signalling goes
down.
1092 asxQ2931Up This trap is generated whenever UNI signalling goes
up.
1093 asxFabricDown An asxFabricDown trap signifies that the sending pro-
tocol entity recognizes a failure in one of the ASX-4000
ATM Switch fabrics, that is identified by the board
number. This is probably caused by a hot-swap of a
fabric module.
1094 asxFabricUp An asxFabricUp trap signifies that the sending proto-
col entity recognizes a new operational ATM ASX-4000
Switch fabric, that is identified by the board number.
This is probably caused by a hot-swap of a fabric mod-
ule.

Configuring SNMP
Trap
Number
1095 asxQ2931CallClearing This trap is generated whenever calls are cleared in a
signalling interface for the following reasons.
-- Carrier loss
-- SONET alarms in case of SONET
-- VP failure
2000 frf8PVCStatus This trap indicates when an interworking PVC has
experienced an alarmed condition, either on the ATM
network side or Frame Relay side. It is also generated
when the PVC alarmed condition is cleared. It carries
the operational status of the PVC by the
frf8ConnOperStatus, as well as the reason why exiting,
entering, or changing the alarmed state
frf8ConnPVCAlarmReason.
If the interworking PVC changes its status by an

AdminStatus command (active/inactive/testing),
causing the frf8ConnOperStatus to change (up/down),
this trap is generated with frf8ConnPVCAlarmReason
set to no defect. This trap gives the support to manage
defects that might occur in either the Frame Relay net-
work or the ATM network, even if the PVC Opera-
tional Status is already down.
2001 ifLinkDown This trap indicates that the SNMP agent has detected
that the ifOperStatus of this interface is about to transi-
tion into the down state.
2002 ifLinkUp This trap indicates that the SNMP agent has detected
that the ifOperStatus of this interface has transitioned
out of the down state.
2003 framNakMsg This trap is sent when a message from the SCP to the
FramePlus network module does not succeed. The
request that has failed is encoded within the trap as the
messageType, and the reason for failure is encoded
within the errorCode. See Table A.2 for the message
types and the message requests. See Table A.3 for the
error codes and their meanings.

Configuring SNMP
Configuring SNMP
Trap
Number
2004 pnniSpvccDown This trap is sent when an SPVC call is cleared or
deleted. The pnniSpvcSrcDownReason indicates if the
call was cleared due to a better route being found or a
network failure. It also may indicate SPVC deletion.
This trap is only sent if pnniSpvcTrapMode is all.
2005 pnniSpvccUp This trap is send when an SPVC call is established. It
identifies the SPVC and the old and new cost. If the
call has not been up before, the pnniSpvcSrcOldRoute-
Cost will be -1. This trap is only sent if pnniSpvcTrap-
Mode is failure or all.
2006 pnniSpvccFail This trap is sent when an SPVC call that is down fails
to get connected on 16 consecutive attempts. This trap
is only sent if pnniSpvcTrapMode is failure or all.
2007 pnniSpvpcDown This trap is sent when an SPVP call is cleared. The
pnniSpvcSrcDownReason indicates if the call was
cleared due to a better route being found or a network
failure. It also may indicate SPVPC deletion. This trap
is only sent if pnniSpvpcTrapMode is all.
2008 pnniSpvpcUp This trap is sent when an SPVP call is established. It
identifies the SPVP and the old and new cost. If the call
has not been up before, the pnniSpvcSrcOldRouteCost
will be -1. This trap is only sent if pnniSpvpcTrapMode
is all or failure.
2009 pnniSpvpcFail This trap is sent when an SPVP call that is down fails
to get connected on 16 consecutive attempts. This trap
is only sent if pnniSpvpcTrapMode is all or failure.
2010 asxPortCardDown An asxPortCardDown trap signifies that the sending
Switch port cards, that is identified by the port card
index. This is probably caused by a hot-swap of the
port card.
2011 asxPortCardUp An asxPortCardUp trap signifies that the sending pro-
tocol entity recognizes a new operational ATM Switch
port card, that is identified by the port card index. This
is probably caused by a hot-swap of the port card.

Configuring SNMP
Trap
Number
2013 asxServiceCategoryOutput- This trap indicates that the output queue for the ser-
QueueCongested vice category has exceeded its dedicated length, and
has begun overflowing into the shared buffer space on
the netmod.
2014 asxServiceCategoryOutput- This trap indicates that the output queue for the given
QueueCellLoss service category has overflowed and cells have been
dropped.
2015 pnniNormalToOver- This trap is generated on transition from the normal
loadTransition state to the overload state.
2016 pnniOverloadToNormal- This trap is generated on transition from the overload
Transition to the normal state.

Configuring SNMP
Table A.2 defines the message types and the message requests for Trap 2003.
Configuring SNMP
Table A.2 - Message Type Encodings for Trap 2003
MessageType Message Request

0 Configure network module
1 Configure port
2 Send egress Fdl
4 Network module reset
5 Port reset
6 Configure a channel/service
7 Request channel/service configuration
8 Request port configuration
9 Request connection configuration
10 Channel/service reset
17 Configure LEDs
28 Request application revision
29 Create a FUNI service
30 Create a Frame Relay service
31 Request to change service admin status
32 Request to enable/disable service statistics
33 Request to enable/disable network module statistics
34 Request to enable/disable network module admin status
35 Request to enable/disable port admin status
36 Delete a service
37 Create a Frame Relay connection
38 Create a FUNI connection
40 Request to enable/disable connection admin status
41 Request to enable/disable service egress rate enforcement

Configuring SNMP
Table A.2 - Message Type Encodings for Trap 2003 (Continued)
MessageType Message Request

42 Request to enable/disable connection ingress rate enforcement
43 Delete a connection
47 Configure port to be timing source
Table A.3 defines the error codes for Trap 2003 and their meanings.
Table A.3 - Error Codes for Trap 2003
Error Code Error Code Meaning

1 Invalid or out of range parameter
2 System disabled
3 Memory allocation failure
4 Timer allocation failure
5 Circuits on link exceeded
6 Invalid CIR configuration
7 No CCB for link
8 Unknown link
9 Unknown DLCI
0A Link number out of range
0B DLCI number out of range
0C Link already exists
0D Circuit on link already exists
0E Link contains circuits
0F Out of memory
10 Out of timers
11 Out of timers
12 Length error
13 Duplicate connection ID

Configuring SNMP
Table A.3 - Error Codes for Trap 2003 (Continued)
Configuring SNMP
Error Code Error Code Meaning
14 Invalid parameter
15 Duplicate DLCI
16 Memory allocation failure
17 Service not created
18 Connection does not exist
19 Invalid Service
1A Connection ID unavailable
1B Invalid Connection ID
1C Unexpected message
1D Object already configured
21 Operation failed
22 Physical channel configuration failed
23 Physical connection configuration failed
24 Physical channel teardown failed
25 Physical connection teardown failed
26 Port configuration failed
27 EPD/PPD configuration failed
28 Unknown SDPM message received
29 Netmod not configured
2A SDPM message size error
2B Port not configured
2C Allocation of buffer failed

Configuring SNMP
A.2.1 Adding SNMP Trap Destinations

To create one or more SNMP trap destinations on a switch, log in to AMI and open a session
on the switch. Enter the following parameters:
system snmp trapdest new -trapdest <IP address>
The <IP address> variable indicates the IP address of the SNMP trap destination that is to
be created. For example:
system snmp trapdest new -trapdest 193.14.22.10
Repeat this for as many SNMP trap destinations as needed. Traps are active as soon as they are
set.
A.2.2 Displaying SNMP Trap Destinations

To list all of the current SNMP trap destinations, log in to AMI and open a session on the
switch. The SNMP traps supported by this switch are detailed in the FORE-Switch-MIB. The
switch MIBs can be found on the TACtics Online web page at:
http://www.fore.com/tac/index.html
To download the MIBS:

1. Log in to TACtics Online.
2. In the left-hand column, choose Software.
3. From the yellow boxes at the top of the screen, choose Mibs.
4. Select the atm-switch product category and choose the ForeThought version that
you would like to download.
To display the SNMP trap destinations, enter the following AMI command on a switch:
system snmp trapdest show
The switch responds with a list similar to the following:
Trap Destination
1 192.88.243.18
2 198.29.16.14
3 198.29.16.18

Configuring SNMP
If no trap destinations have been configured, then the following is displayed:
Configuring SNMP
myswitch:system snmp trapdest-> show
No trap information is available
A.2.3 Removing SNMP Trap Destinations

To delete one or more SNMP trap destinations for a switch, log in to AMI and open a session
on the switch. Prior to deleting any trap that may need to be recreated later, as a precaution, a
recommended practice is to list all trap destinations using AMI and either copy the screen or
write down the destinations. To delete a trap, enter the following command:
system snmp trapdest delete -trapdest <IP address>
The <IP address> variable indicates the index number of the SNMP trap destination that is
to be removed. For example:
system snmp trapdest delete -trapdest 193.14.22.10
Repeat this for as many SNMP trap destinations as needed.
For more information about the SNMP trap

NOTE destination commands, see Part 3 of the AMI

Configuring SNMP

APPENDIX B Configuring Circuit Emulation
Services
FORE Systems’ Circuit Emulation Services (CES) Network Modules (NMCE-6/DS1A and
NMCE-6/E1A) provide adaptation from time-division multiplexed (TDM) equipment (i.e.,
PBXs, WAN multiplexers, channel banks, video codecs, etc.) and traffic to ATM. Both modules
Emulation Services
Configuring Circuit
provide structured and unstructured services, with a maximum of 127 connections supported
on each module.
• All six ports of the DS1 CES Network Module may support fractional DS1
services (n x 56 Kbps/n x 64 Kbps) where 1 to 24 contiguous or non-contiguous
DS0 channels are mapped to a single ATM VCC, not to exceed 127 total connec-
tions.
• All six ports of the E1 CES Network Module may support fractional E1 services (n
x 56 Kbps/n x 64 Kbps) where 1 to 31 contiguous or non-contiguous DS0 channels
are mapped to a single ATM VCC, not to exceed 127 total connections.
Structured services provide digital access and cross-connect system (DACS) connectivity
where n x 64 Kbps and n x 56 Kbps digital signal level zero (DS0) channels are adapted to
ATM cells and mapped to unique ATM virtual connections (VCCs).
Unstructured services provide support and maintenance of a single full bandwidth 1.544
Mbps (DS1) or 2.048 Mbps (E1) clear channel across a single ATM virtual connection.
Configurations of both the DS1 and the E1 version of the CES network module are detailed in
this chapter. This chapter assumes the proper configuration of general switch parameters and
of CES specific parameters, such as timing distribution/recovery, additional CES alarms, CES
ports, etc.
ATM Switch Network Configuration Manual B-1

Configuring Circuit Emulation Services
B.1 Idle Channel Suppression

Idle channel suppression is a method used to detect active and inactive timeslots (data paths
of size DS0) in a TDM trunk being carried over an ATM network based on CES. An idle
timeslot is a timeslot in a connection through which no data is transmitted; e.g., that timeslot
may not be carrying an active connection. An active connection is one in which at least one
timeslot is not idle, and an inactive connection is one in which all of the timeslots are idle.
Because the CES services are CBR services, when a CES connection is created, the switch allo-
cates a fixed amount of bandwidth, depending on the number of timeslots used. When idle
channel suppression is enabled on a CES connection, the CES network module transmits only
the cells of active connections to the switch fabric. Thus, only a fraction of the bandwidth allo-
cated for the connection is used. The egress network module detects that there is less than the
contracted cell rate arriving on the CES connections, and re-allocates the extra bandwidth to
other existing VBR, ABR, and UBR connections, but does not exceed their contracted cell rates.
Additionally, the egress network module guarantees the constant bit rate for CES connections
whenever they want to reclaim the bandwidth.
When CES SPVCs or PVCs are created, the virtual channels assigned for these connections at
the intermediate switches are also assigned CBR service. The unused bandwidth in the interme-
diate connections are also reused by other bit rate services. Idle channel suppression works
slightly differently for Channel Associated Signalling (CAS) connections than it does for basic
connections.
B.1.1 CAS Connections

The timeslot state is determined according to CAS. When all of the timeslots on a given connec-
tion become idle, the connection becomes idle. A short period of time (the idle integration period)
is waited, during which the idle patterns are observed, before the channel state is changed from
active to idle. When a connection is idle, OAM F5 cells are used to maintain the connection. When
one or more channels on a given connection become active, the connection becomes active. When
an active connection is detected, the change from idle to active is immediate.
B.1.1.1 Idle Detection using Signalling Bits

This method can be used only for structured CAS connections and is a recommended method
for voice connections. The majority of these applications use the AB bit idle state patterns. The
CAS patterns supported are as follows: E&M, Loop Start FXS, Loop Start FX0, Ground Start
FXS, and Ground Start FX0. These patterns can be user-to-network or network-to-network.
The available options are as follows: em00, em01, fxolsuser01, fxolsuser11, fxolsnet00,
fxolsnet01, fxslsuser00, fxslsuser01, fxslsnet01, fxslsnet11, fxsgsuser01, fxsgsnet10, fxsgsnet11,
fxogsuser10, fxogsuser11, fxogsnet01, r210. The default pattern is em00. The default integra-
tion period is 1 second and is user-configurable from 500 milliseconds to 2 seconds.
B-2 ATM Switch Network Configuration Manual

B.1.1.2 Active Detection

A connection is declared active when at least one timeslot becomes active. The active detection
mechanism is not user-configurable. The AB bits are compared against the configured pattern
for idle states, and if different, the connection is immediately declared active.
B.1.2 Basic Connections

For Basic (Common Channel Signalling (CCS)) connections, only one idle pattern is used to
determine the connection state. When the data received from the line interface matches the
idle pattern for a short period of time (the integration period), the connection becomes idle.
Emulation Services
Configuring Circuit
When a connection is idle, OAM F5 cells are used to maintain the connection. When the first
data cell that is received does not match the idle pattern, the connection immediately becomes
active.
Idle channel suppression is not supported on

NOTE unstructured connections.
Structured basic connections require a partial fill

size of at least 24 to support idle channel
suppression.
B.1.2.1 Idle Detection using Pattern Matching

This method of idle detection is used only for basic (CCS) services. Depending on the applica-
tions, the idle pattern can vary. Additionally, different idle patterns can be used to identify, for
example, an out-of-service channel.
In this implementation, a single idle pattern determines the idle status of a channel. The
default idle pattern is FF.
An idle code pattern must be consistent during an idle integration period before it can be
determined that a time slot has become idle. This eliminates the false detection of an idle state
due to the natural occurrence of the specified idle data patterns in the connection’s normal
data content from time to time. For basic connections, the default integration period is 1 sec-
ond, and is user-configurable from 500 milliseconds to 2 seconds.
B.1.2.2 Idle Detection using Mask Patterns

This method of idle detection is used only for basic (CCS) services. The idle patterns can be
provided by the user in a mask format. The mask pattern is provided in a format, for example,
F0 hex. This means that pattern matching is performed only on the four most significant bits.
This method is useful, for example, in T1 robbed-bit signalling systems.

B.1.2.3 Active Detection

A connection is declared active when at least one timeslot becomes active. The active detection
mechanism is not user-configurable. If at least one of the timeslots is not matched against the
idle patterns configured, the connection is declared active. The integration period is not user-
configurable and the transition from idle to active is immediate; i.e., when the first cell does
not have idle patterns.
B.1.3 Idle Channel Suppression Configuration Examples

Various examples are provided in this section for configuring idle channel suppression.
For proper operation, idle channel suppression

NOTE must be enabled on both the input and the
output side of the connection.
B.1.3.1 Enabling Idle Channel Suppression

Idle channel suppression is enabled on a per-connection basis. The default idle detection
mechanism is idle pattern matching. To enable idle channel suppression, you could use the
following syntax if both the input and output ports are CEM ports that are directly connected:
myswitch:services ces-> new -iatmif <AtmIf> -its <text> -oatmif <AtmIf> -ots <text>
-idlesupp <state> -oidlesupp <state>
For example:
myswitch:services ces-> new -iatmif 1a3 -its 1 -oatmif 1b1 -ots 1 -idlesupp enable
-oidlesupp enable
Alternatively, you could use the following syntax if the the input and output ports are CEM
ports that are not directly connected (e.g., an OC-3 port in between). Enter the following on
the incoming switch:
incomingswitch:services ces-> new -iatmif <AtmIf> -its <text> -oatmif <AtmIf> -ovpi
<integer> -ovci <integer> -idlesupp <state>
Enter the following on the outgoing switch:
outgoingswitch:services ces-> new -iatmif <AtmIf> -its <text> -oatmif <AtmIf> -ovpi
<integer> -ovci <integer> -oidlesupp <state>

For example:
incomingswitch:services ces-> new -iatmif 1a3 -its 1 -oatmif 1b1 -ovpi 0 -ovci 100
-idlesupp enable
outgoingswitch:services ces-> new -iatmif 1a3 -its 1 -oatmif 1b1 -ovpi 0 -ovci 100
-oidlesupp enable
B.1.3.2 Configuring Idle Detection

Idle channel suppression is enabled when any idle channel suppression option is used in the
Emulation Services
Configuring Circuit
AMI command line. An idle detection mechanism is used on a per-connection basis. When
idle pattern matching is used, the default pattern is FF. The same default pattern is used for
mask pattern matching. When the signalling method is used, the default value is as per Cus-
tomer Installation to Network idle patterns. The idle detection method is based on the connec-
tion type as follows:
• CAS connections use idle CAS signalling pattern matching.
• Structured Basic connections use idle pattern matching.
B.1.3.2.1 Configuring Idle CAS Signalling Pattern Matching

Standard idle signalling patterns are specified by name. Only one type can be selected. To con-
figure CAS signalling pattern matching, you could use the following syntax:
myswitch:services ces-> new -iatmif <AtmIf> -its <text> -idlesupp

<state> -idlepat <text> -oatmif <AtmIf> -ots <text>
-oidlesupp <state> -oidlepat <text> -cas <cas>
For example:
myswitch:services ces-> new -iatmif 1a3 -its 1 -idlesupp enable -idlepat

fxsluser01 -oatmif 1b2 -ots 1 -oidlesupp enable -oidlepat em01 -cas cas
Alternatively, you could use the following syntax:
myswitch:services ces-> new -iatmif <AtmIf> -its <text> -idlesupp <state> -idlepat
<text> -oatmif <AtmIf> -ovpi <integer> -ovci <integer> -cas <cas>
For example:
myswitch:services ces-> new -iatmif 1a3 -its 1 -idlesupp enable -idlepat em01
-oatmif 1b2 -ovpi 0 -ovci 100 -cas cas

B.1.3.2.2 Configuring an Idle Pattern for Pattern Matching

If an idle pattern is configured, by default, idle channel suppression and idle suppression
based on pattern matching are enabled. The idle pattern is used on per-connection basis. A
maximum of one pattern can be used for structured basic connections. To configure an idle
pattern for pattern matching, you could use the following syntax:

<state> -idlepat <text> -oatmif <AtmIf> -ots <text> -oidlesupp
<state> -oidlepat <text>
For example:
myswitch:services ces-> new -iatmif 1a3 -its 1 -idlesupp enable -idlepat 1F

-oatmif 1b2 -ots 1 -oidlesupp enable -oidlepat 2F

<state> -idlepat <text> -oatmif <AtmIf> -ovpi <integer> -ovci <integer>
For example:

-oatmif 1b2 -ovpi 0 -ovci 100
B.1.3.2.3 Configuring Mask Patterns

To configure mask patterns, you could use the following syntax:

<state> -idlemask <integer> -idlepat <text> -oatmif <AtmIf>
-ots <text> -oidlesupp <state> -oidlemask <integer> -oidlepat <text>
For example:
myswitch:services ces-> new -iatmif 1a3 -its 1 -idlesupp enable -idlemask F0

-idlepat 2F -oatmif 1b2 -ots 1 -oidlesupp enable -oidlemask F0 -oidlepat 2F


<state> -idlemask <integer> -idlepat <text> -oatmif <AtmIf>
-ovpi <integer> -ovci <integer>
For example:
myswitch:services ces-> new -iatmif 1a3 -its 1 -idlesupp enable -idlemask FF
Emulation Services
Configuring Circuit
-idlepat 7F -oatmif 1b2 -ovpi 0 -ovci 100
B.1.3.3 Configuring an Idle Integration Period

The idle integration period is provided in milliseconds. Only an active to idle integration
period is supported. To configure the idle integration period, you could use the following syn-
tax:

<state> -idlepat <text> -idleintp <integer> -oatmif <AtmIf>
-ots <text> -oidlesupp <state> -oidlepat <text> -oidleintp <integer>
For example:

-idleintp 500 -oatmif 1b2 -ots 1 -oidlesupp enable -oidlepat 1F -oidleintp 600

<state> -idlepat <text> -idleintp <integer> -oatmif <AtmIf>
-ovpi <integer> -ovci <integer>
For example:

-idleintp 500 -oatmif 1b2 -ovpi 0 -ovci 100

B.2 Configuring CES Connections

The ces commands let you create and delete CES connections, as well as display the status of
existing connections. You can display the list of available subcommands by typing ? at the
ces level.
myswitch:services ces-> ?
delete Delete a CES configuration
new Create a new CES configuration
show Display CES configuration
B.2.1 Deleting a CES Connection

To delete a CES connection, enter the following parameters:
myswitch:services ces-> delete -ifindex <integer>
For example:
myswitch:services ces-> delete -ifindex 20031
The parameter for delete is defined as follows:
service The CES service ID of the connection to be deleted.

B.2.2 Creating a New CES Connection

To create a new CES connection, you must set some of the following parameters:
myswitch:services ces-> new

Usage:
[[-ifindex] <integer>] CES Service Index
[-its] <text> Input Timeslots
[[-oatmif] <AtmIf>] Output AtmIf
Emulation Services
Configuring Circuit
[[-ots] <text>] Output Timeslots
[[-ovpi] <integer>] Output VPI
[[-ovci] <integer>] Output VCI
[[-srts] <cbrclockmode>] SRTS
[[-fupc] <UPC Index>] Forward UPC
[[-bupc] <UPC Index>] Backward UPC
[[-cas] <cas>] CAS
[[-partialfill] <integer>] Partial Fill
[[-reassCDVT] <integer>] Reass CDVT (us)
[[-bufSize] <integer>] Buf Size
[[-integ] <integer>] Integer
[[-idlesupp] <state>] Idle Supp
[[-idlemask] <Hex Integer>] Idle Mask
[[-idlepat] <text>] Idle Pat
[[-idleintp] <integer>] Idle Int Period
[[-oidlesupp] <state>] Oidle Supp
[[-oidlemask] <Hex Integer>] Oidle Mask
[[-oidlepat] <text>] Oidle Pat
[[-oidleintp] <integer>] Oidle Intp
Enter the following to create a new CES connection:
myswitch:services ces-> new -iatmif <AtmIf> -its <text>
For example:
myswitch:services ces-> new -iatmif 1a3 -its 1

The CES new command can also be used in other ways. When the following parameters are
used, by default, a bidirectional PVC is created.
myswitch:services ces-> new -iatmif <AtmIf> -its <text> -oatmif <AtmIf> -ots
<text>
OR
myswitch:services ces-> new -iatmif <AtmIf> -its <text> -oatmif <AtmIf> -ovpi
<integer> -ovci <integer>
For example:
myswitch:services ces-> new -iatmif 1a3 -its 2 -oatmif 3b1 -ots 2
OR
myswitch:services ces-> new -iatmif 1a1 -its 1 -oatmif 1d1 -ovpi 0 -ovci 100
SRTS is only available on unstructured

NOTE connections, which are created by specifying
all for the -its <text> or -ots <text>
parameter.
The -cas, -partialfill, -idlesupp, and

-oidlesupp options are not applicable to
unstructured mode.
Structured mode is selected by indicating the

exact timeslots to be used. For example, timeslots
1, 2, and 3 would be entered as 1-3, timeslots 2, 4,
and 6 would be entered as 2:4:6, and
combinations such as 1-4:9-11:12 are allowed.
The parameters for new are defined as follows:
-ifindex <integer> The index number to be assigned to this CES service. An index is automatically assigned
by the switch. The default is 0.
-iatmif <AtmIf> The input port on which the CES connection is to be created.
-its <text> Indicates which timeslots (1-24 for DS1, 1-31 for E1) are being configured for a particular
PVC. all indicates unstructured service. The time slot assignments may be either contig-
uous or non-contiguous DS0s.
B - 10 ATM Switch Network Configuration Manual

-oatmif <AtmIf> The output port of the CES connection, which can be a CES port or an ATM port.
-ovpi <integer> The output Virtual Path Identifier (VPI) of the CES connection when the output port is not
a CES port.
-ovci <integer> The output Virtual Channel Identifier (VCI) of the CES connection when the output port is
not a CES port.
-ots <text> The output timeslots of the CES connection when the output port is a CES port.
-srts <cbrclockmode> Indicates whether Synchronous Residual Time Stamp (SRTS) clock recovery is to be
enabled on this connection. srts indicates that SRTS is enabled. synch indicates that
Emulation Services
Configuring Circuit
SRTS is disabled. The default is synch.
-fupc <UPC Index> The UPC contract type to be used in the ingress direction of the connection. (See Part 1 of
the AMI Configuration Commands Reference Manual for more information about UPC con-
tracts.)
-bupc <UPC Index> The UPC contract type to be used in the egress direction of the connection. (See Part 1 of
tracts.)
-cas <cas> Indicates whether Channel Associated Signalling (CAS) is to be used on the connection.
basic indicates that CAS will not be used, cas indicates that CAS will be used. The
default is basic.
-partialfill <integer> Indicates how many of the available 47 payload bytes in each cell are used before they are
deemed “full” and ready for transmission across the ATM network (i.e., how much of the
ATM cell contains data and how much is padding). The range for this parameter is 12 to
47. The default value is 47. Partialfill is used to minimize network transmission latency
and is useful especially with time-sensitive, robbed-bit signalling sources.
-reassCDVT <integer> The Cell Delay Variation Tolerance for cells being received by the segmentation and reas-
sembly (SAR) engine. This value must be entered as a multiple of 10 (µs). The range for
this parameter is 100 to 24000 (in µs), and the default is 2000.
-bufSize <integer> The amount of reassembly buffer space allocated for the connection. The default is 256
bytes per timeslot.
-integ <integer> The amount of time allocated to re-establish the connection before, while, or after the call
is established, or in the case of interruption. The default is 2500 milliseconds.
-idlesupp <state> enable means idle channel suppression is going to be used on the incoming side of this
connection. disable means idle channel suppression is not going to be used on the
incoming side of this connection. The default is disable. When this option is enabled on
structured basic connections, a minimum partialfill size of 24 is required. For proper oper-
ation, idle channel suppression must be enabled on both the input and the output side of
the connection.
-idlemask <integer> The mask pattern for idle detection on the incoming side of this connection. This method
is only used for structured basic service. The range of values is 00 to FF. The default is FF.
ATM Switch Network Configuration Manual B - 11

a CES port.
not a CES port.
tracts.)
tracts.)
default is basic.
bytes per timeslot.
the connection.

a CES port.
not a CES port.
Emulation Services
Configuring Circuit
tracts.)
tracts.)
default is basic.
bytes per timeslot.
the connection.

-idlepat <text> The pattern for idle detection on the incoming side of this connection. For detection based
on both idle and mask patterns, it contains the idle octet pattern. For detection based on
signalling, it contains one of the following CAS patterns: em00, em01, fxolsuser01,
fxolsuser11, fxolsnet00, fxolsnet01, fxslsuser00, fxslsuser01, fxslsnet01, fxslsnet11,
fxsgsuser01, fxsgsnet10, fxsgsnet11, fxogsuser10, fxogsuser11, fxogsnet01, r210. A maxi-
mum of one idle pattern can be used for structured basic connections. Idle patterns are
filled from the least significant byte. The default is FF for structured basic connections. The
default CAS idle AB bit pattern is em00.
-idleintp <integer> The integration period for idle detection on the incoming side of this connection. Idle pat-
terns are observed for this period before declaring that an active connection has gone idle.
For both CAS and basic connections, the range of values is 500 milliseconds to 2 seconds
and the default is 1 second.
-oidlesupp <state> enable means idle channel suppression is going to be used on the outgoing side of this
connection. disable means idle channel suppression is not going to be used on the out-
going side of this connection. The default is disable. When this option is enabled on
the connection.
-oidlemask <integer> The mask pattern for idle detection on the outgoing side of this connection. This method is
only used for structured basic service. The range of values is 00 to FF. The default is FF.
-oidlepat <text> The patterns for idle detection on the outgoing side of this connection. For detection based
on both idle and mask patterns, it contains the idle octet patterns. For detection based on
signalling, it contains one of the following CAS patterns: em00, em01, fxolsuser01,
fxolsuser11, fxolsnet00, fxolsnet01, fxslsuser00, fxslsuser01, fxslsnet01, fxslsnet11,
fxsgsuser01, fxsgsnet10, fxsgsnet11, fxogsuser10, fxogsuser11, fxogsnet01, r210. A maxi-
mum of one idle pattern can be used for structured basic connections. Idle patterns are
filled from the least significant byte. The default is FF for structured basic connections. The
default CAS idle AB bit pattern is em00.
-oidleintp <integer> The integration period for idle detection on the outgoing side of this connection. Idle pat-
terns are observed for this period before declaring that an active connection has gone idle.
For both CAS and basic connections, the range of values is 500 milliseconds to 2 seconds
and the default is 1 second.

B.2.3 Displaying CES Connections

To display the current CES connections, enter the following:
myswitch:services ces-> show

CES Input Output
Index State AtmIf TimeSlots VPI VCI Type AtmIf TimeSlots VPI VCI
1038 down 4A1 1 0 129 pvc 4A4 2 0 130
1050 down 4A4 2 0 130 pvc 4A1 1 0 129
1051 down 4A2 3 0 131 pvc 4A3 4 0 132
Emulation Services
Configuring Circuit
1052 down 4A3 4 0 132 pvc 4A2 3 0 131
Field Description
CES Index The identification number (assigned by the switch) of this CES connection.
Input AtmIf The incoming port on which the CES connection exists.
Input State Indicates whether the CES connection is enabled (up) or disabled (down).
Input Timeslots Indicates which timeslots (1-24 for DS1, 1-31 for E1) are configured for the input port. all
indicates unstructured service.
Input VCI The incoming VCI value of the connection.
Input VPI The incoming VPI value of the connection.
Output AtmIf The outgoing port on which the CES connection exists.
Output Timeslots Indicates which timeslots (1-24 for DS1, 1-31 for E1) are configured for the output port.
all indicates unstructured service.
Output VCI The outgoing VCI value of the connection.
Output VPI The outgoing VPI value of the connection.
Type The type of ATM connection (i.e., PVC or SPVC) that is associated with the CES connec-
tion.

B.2.3.1 Displaying Advanced Information

To display the advanced settings of the current CES connections, enter the following:
myswitch:services ces-> show -advanced

CES Map Map Service Clock Partial Max Integ. reass
Index VPI VCI Type Mode Cas Fill Buf Period CDVT
1038 0 129 structured synch basic 47 256 2500 2000
Field Description
MapVPI The incoming VPI value of the connection.
MapVCI The incoming VCI value of the connection.
Service Type Shows if this connection uses structured or unstructured service.
Clock Mode Synch means that the connection is in synchronous mode (either structured or unstruc-
tured). SRTS means that the connection is in asynchronous (unstructured) mode. (Syn-
chronous Residual Time Stamp (SRTS) clock recovery is enabled on this connection.)
Cas basic indicates that Channel Associated Signalling (CAS) will not be used, cas indicates
that CAS will be used.
Partial Fill Indicates how many of the available 47 payload bytes in each cell are used before they are
ATM cell contains data and how much is filler). partialfill is used to minimize network
transmission latency and is useful especially with time-sensitive, robbed-bit signalling
sources.
Max Buf The amount of reassembly buffer space allocated for the connection. The default is 256
bytes per timeslot.
Integ. Period The amount of time allocated to re-establish the connection before, while, or after the call
reass CDVT The Cell Delay Variation Tolerance for cells being received by the segmentation and reas-
sembly (SAR) engine.

B.2.3.2 Displaying Idle Channel Suppression Information

To display the idle channel suppression information for the current CES connections, enter the
following:
myswitch:services ces-> show -idlesuppression

CES IdleSupp Idle Idle NoOf Idle IdleInt.
Index State Type Mask Patt Patterns Period(ms)
1038 enabled pattern 255 1 255 1000
Emulation Services
Configuring Circuit
Field Description
IdleSupp State enabled means idle channel suppression is going to be used on this connection.
disabled means idle channel suppression is not going to be used on this connection.
Idle Type The idle detection mechanism used on this connection.

pattern means idle detection is based on idle patterns.
cas means idle detection is based on CAS signalling. This field is read-only.
Idle Mask The mask pattern for idle detection on this connection. This method of idle detection can
only be used on basic and unstructured connections.
NoOf Patt The number of idle patterns configured for detection. This field is read-only.
Idle Patterns The patterns for idle detection on this connection. For detection based on both idle and
mask patterns, it contains the idle octet patterns. For detection based on signalling, it con-
tains one of the following patterns: em00, em01, fxolsuser01, fxolsuser11, fxolsnet00,
fxolsnet01, fxslsuser00, fxslsuser01, fxslsnet01, fxslsnet11, fxsgsuser01, fxsgsnet10,
fxsgsnet11, fxogsuser10, fxogsuser11, fxogsnet01, r210. A maximum of one idle pattern can
be used for structured basic connections. Idle patterns are filled from the least significant
byte.
IdleInt. Period (ms) The integration period for idle detection on this connection. Idle patterns are observed for
this period before declaring that an active connection has gone idle.

B.2.4 Displaying CES Statistics

To display the current CES statistics, enter the following:
myswitch:services ces-> statistics

CES Reass Header Pointer Buffer Buffer CellLoss
Id Cells Errors Reframes Lost Cells UnderFlows OverFlows Status
1061 13911864 0 0 0 0 6 noLoss
1062 5883 0 0 0 0 6 noLoss
Field Description
CES Id The CES service ID of the CES connection for which statistics are shown.
Reass Cells The number of cells that have been reassembled on the port.
Header Errors The number of AAL1 header errors that have been seen on the port.
Pointer Reframes The number of errors encountered in the AAL1 pointer.
Lost Cells The number of cells that have been lost on the port.
Buffer UnderFlows The number of bytes lost due to missing cells.
Buffer OverFlows The number of extra bytes received which were not anticipated.
CellLoss Status Indicates whether the AAL1 state machine is currently in a state where cells are not being
received. Displays one of the following: loss, no loss, idle, or no cells.

APPENDIX C Converting FT-PNNI and PNNI
Networks
This appendix discusses the conversion of both non-hierarchical and hierarchical FT-PNNI
networks to ATM Forum PNNI (hereafter referred to as PNNI) routing. It also describes the
conversion of a non-hierarchical PNNI network to a lop-sided hierarchical PNNI network. It is
assumed that you are familiar with the fundamentals of FT-PNNI and PNNI routing and
familiar with FORE’s implementation of PNNI as described in Chapter 6 of this manual.
The first section discusses PNNI routing in networks that contain ASX-1000, ASX-1200, or
TNX-1100 switches. The various procedures for converting your network are described in the
later sections. If you do not have any ASX-1000, ASX-1200, or TNX-1100 switches in your net-
work, you can skip to Section C.2 or Section C.3 to learn how to convert your network from
FT-PNNI to PNNI. You can skip to Section C.4 to learn how to convert your non-hierarchical
PNNI network to a lop-sided hierarchical PNNI network. If you do have ASX-1000, ASX-1200,
or TNX-1100 switches in your network, it is recommended that you read Section C.1 first to
Converting FT-PNNI
and PNNI Networks
avoid potential configuration problems when migrating your network.
• Section C.1 - ASX-1000, ASX-1200, and TNX-1100 Routing Configuration Issues
• Section C.2 - Migration of a Non-Hierarchical FT-PNNI Network
• Section C.3 - Migration of a Hierarchical FT-PNNI Network
• Section C.4 - Migration of a Non-hierarchical PNNI Network to a Lop-sided
Hierarchical PNNI Network
ATM Switch Network Configuration Manual C-1

Converting FT-PNNI and PNNI Networks
C.1 ASX-1000, ASX-1200, and TNX-1100 Routing

Configuration Issues
This section discusses routing in hierarchical networks in which there are ASX-1000,
ASX-1200, or TNX-1100 switches present in your network. Because each one of the four fabrics
appears as a single node in routing and because you cannot take a two-hop path to go from
one switch fabric to another switch fabric in the same ASX-1000, ASX-1200, or TNX-1100, there
are certain FT-PNNI and PNNI node configurations that should be avoided. These special con-
figurations are addressed individually and alternate configurations are suggested.
C.1.1 ASX-1000s, ASX-1200s, or TNX-1100s in FT-PNNI Peer

Groups
Figure C.1 shows an invalid configuration of an ASX-1000, ASX-1200, or TNX-1100 divided
between two FT-PNNI peer groups called A and B. The ASX-1000’s, ASX-1200’s, or
TNX-1100’s four FT-PNNI nodes are A.1 and A.2 in peer group A and B.1 and B.2 in peer
group B. From peer group A, the intra-peer group links of peer group B are not visible. So, the
link B.1 to B.2 is not visible in peer group A. Similarly, non-border nodes B.3 and B.4 belonging
to peer group B are not visible in peer group A.
FT-PNNI peer group boundary

FT-PNNI node
FT-PNNI link A.3
A.1 B.1 B.3
Backbone link
PG A PG B
A.5
A.2 B.2
A.4 B.4
Figure C.1 - Invalid Configuration of ASX-1000, ASX-1200, or TNX-1100 Split between Two
FT-PNNI Peer Groups
C-2 ATM Switch Network Configuration Manual

Suppose you want to route from A.4 to B.3. After the initial step of prefix matching, the Peer
Group Summary Node (PGSN) of peer group B (not shown in the figure) is chosen as the des-
tination for this path computation. Since B.1 and B.2 both advertise links to the PGSN and
they are equidistant from A.4, A.4 may decide to construct the Destination Transit List (DTL)
as A.4 to A.2 to B.2 to B’s PGSN. When the setup of this call proceeds into peer group B and
reaches node B.2, the only way to route to B.3 is through B.1. But since you cannot take a two-
hop path across the ASX-1000, ASX-1200, or TNX-1100, this call setup will fail.
This problem can be avoided by not breaking up an ASX-1000, ASX-1200, or TNX-1100
between multiple FT-PNNI peer groups. For example, in the network in Figure C.2, the entire
ASX-1000, ASX-1200, or TNX-1100 could have been made part of peer group B with only A.3,
A.4, and A.5 in peer group A. A.3 and A.4 would be the border nodes in peer group A con-
nected to A.1 and A.2 (re-numbered to have B as their peer group ID) in peer group B.
Converting FT-PNNI
and PNNI Networks

C.1.2 ASX-1000s, ASX-1200s, or TNX-1100s in PNNI Areas

The problem described in Section C.1.1 can also occur in a network of two PNNI peer groups
(areas). For example, in the poorly-constructed network shown in Figure C.2, the fabric on the
top left (fabric 1) is configured as a split switch with two nodes, one in peer group A and the
other in peer group B. All of the other fabrics have one node each. The node on the bottom left
fabric (fabric 2) is in peer group A and the nodes in the other two fabrics (3 and 4) are in peer
group B. The links between the fabric 2 and the fabrics 3 and 4 do not come up because the
peer group IDs of the PNNI nodes connected these links do not match.
This example assumes that there is no PNNI

NOTE Logical Group Node (LGN) hierarchy
configured in this network.

PNNI node
PNNI link
B.1 B.3
A.1
PG A PG B
B.4
A.2
Figure C.2 - Invalid Configuration of ASX-1000, ASX-1200, or TNX-1100 Split between Two PNNI
Peer Groups
The problem in the network shown above occurs when computing a path from A.2 to B.3. In
this case, routing is forced to use a two-hop path across the backplane, thereby failing to set up
a call. To avoid this problem, be sure that you configure all fabrics in an ASX-1000, ASX-1200,
or TNX-1100 as part of the same peer group.

C.1.3 Multiple Gateways in an ASX-1000, or TNX-1100

Another potential problem can occur when an ASX-1000, ASX-1200, or TNX-1100 in a
FT-PNNI area has gateways (split-switches with one FT-PNNI and one PNNI node) to a PNNI
area. The ASX-1000, ASX-1200, or TNX-1100 can only have, at most, one of its fabrics as a gate-
way between the FT-PNNI area and the PNNI area. This is because if two or more of the fab-
rics have gateways to the PNNI area, then the link connecting any two of the gateways will
come up as a PNNI link in the PNNI area. This will divide up the backplane links between
two different peer groups. This invalid configuration is shown in Figure C.3.

PNNI node B.1 B.3
PNNI link
A.1
FT-PNNI node
FT-PNNI link
A.2
B.2 B.4
Converting FT-PNNI
and PNNI Networks
Figure C.3 - Multiple Fabrics of an ASX-1000, ASX-1200, or TNX-1100 Incorrectly Configured as
Gateways
In the above figure, the ASX-1000, ASX-1200, or TNX-1100 shown is predominantly part of the
FT-PNNI peer group B. However, the ASX-1000, ASX-1200, or TNX-1100 has gateways config-
ured to the PNNI peer group A. Since fabric 1 and fabric 2 are both configured as gateways, the
link between them comes up as a PNNI link in peer group A. Because of this, the backplane
links of the ASX-1000, ASX-1200, or TNX-1100 are divided between peer group A and peer
group B. This may cause two-hop paths to be taken across the backplane links and is an invalid
configuration. To avoid this problem, be sure that you configure either only one of the fabrics
on an ASX-1000, ASX-1200, or TNX-1100 as a gateway between a single FT-PNNI area and a
single PNNI area, or all of the fabrics on an ASX-1000, ASX-1200, or TNX-1100 as a gateway
between the same FT-PNNI area and the same PNNI area.
C.1.4 Migrating from FT-PNNI to PNNI Routing

The migration of FT-PNNI networks to PNNI networks is discussed in the following sections.
It is possible that while migrating a network containing ASX-1000, ASX-1200, or TNX-1100
switches, invalid configurations similar to the ones described in this section may occur during
the intermediate steps. To avoid these potential problems, migrate all of the fabrics in an
ASX-1000, ASX-1200, or TNX-1100 at once, keeping them in the same peer groups or areas.

C.2 Migration of a Non-Hierarchical FT-PNNI Network

This section discusses the conversion of a non-hierarchical FT-PNNI network to a non-hierar-
chical (single peer group) PNNI network. It is assumed that all switches in your network are
FORE switches.
C.2.1 Migration Overview

The following basic steps are involved in converting your network. Each of these steps is
described in detail in the following section. It is recommended that you read the entire section
before attempting to change over your network.
1. Upgrade each switch in the network to ForeThought 6.0.x.
2. Choose a switch that is at the edge of the network. Convert it to a gateway switch
by changing the default protocol of this switch’s default domain to gateway and
reboot the switch. Upon rebooting, this switch will come up with an FT-PNNI node
and a PNNI node. The PNNI node is isolated at this point and does not have any
links attached to it.
3. Choose a switch adjacent to the gateway switch, convert it to a gateway switch by
changing its default protocol for the default domain to gateway, and reboot this
switch. It will come up with a FT-PNNI node and a PNNI node. The link between
the two gateway switches is attached on either end to the respective PNNI nodes.
This link is now the first PNNI link in the network. This single link constitutes a
PNNI area connected by the two gateway switches to the FT-PNNI area. The
FT-PNNI area now contains all the links in the network except the one link
between the two gateways. (It is assumed that there is only one link between the
two gateway switches. If there are multiple links, then all of them will become
PNNI links after this step).
4. Choose another switch that is adjacent to either one of the two switches already
converted to gateway and repeat step step 3. As more and more switches are con-
verted to be gateways, the PNNI area becomes progressively larger and the FT-
PNNI area becomes smaller.
5. For each gateway, when the last link gets converted from FT-PNNI to PNNI (i.e.,
when the last FT-PNNI switch directly connected to this switch becomes PNNI),
modify the default protocol of the default domain to PNNI and reboot the switch.
By doing this, the FT-PNNI node on this gateway, which no longer has any link,
will be deactivated.

C.2.2 Detailed Migration Example

This section gives a detailed example of how to change the example non-hierarchical FT-PNNI
network shown in Figure C.4 to a PNNI network.
FT-PNNI node
FT-PNNI link S2 S3
S1
S4 S5
Figure C.4 - A Non-Hierarchical FT-PNNI Network
This network has five switches named S1 through S5. The following steps are involved in
changing over this network.
Converting FT-PNNI
and PNNI Networks
1. Upgrade all of the switches to ForeThought 6.1.x using the following AMI
command:
For more information about upgrading your switches, see Chapter 4 of the ATM
Switch Installation and Maintenance Manual.
2. Convert S5 to a gateway switch by modifying the default protocol of the default
domain in S5 to gateway using the following AMI command:
routing domain modify -id <integer> -defaultproto gateway
Reboot switch S5. Upon the reboot, S5 will come up with two nodes: one FT-PNNI
node and one PNNI node. By default, these nodes are in areas 4 and 5, respectively,
and levels 4 and 5, respectively. At this point, the nodes in the network are divided
between the two areas with the PNNI node being the only node in area 5 as shown
in Figure C.5.


PNNI node S2 S3
FT-PNNI node
FT-PNNI link
S1
S5
Fore Area A4
Fore Level L4 A5
S4 L5
Figure C.5 - S5 Changed to a Gateway Switch

Reboot switch S3. Upon the reboot, S3 will come up with two nodes: one FT-PNNI
node in area 4 and one PNNI node in area 5. The link between the two gateway
switches S5 and S3 will come up as a PNNI link as shown in Figure C.6. This is the
first PNNI link in the network. All of the other links are still FT-PNNI links.

PNNI node Fore Area A4
PNNI link Fore Level L4
FT-PNNI peer group boundary S3
FT-PNNI node S2
A5
FT-PNNI link L5
S1
S4 S5
Figure C.6 - S3 Changed to a Gateway Switch


Reboot switch S4. Upon the reboot, S4 will come up with a FT-PNNI node in area
4 and a PNNI node in area 5. The link between S4 and S5 will come up as a PNNI
link, and the link between S3 and S4 will come up as a PNNI link, since they are
both gateway switches.
5. At this point, S5 no longer has any links attached to its FT-PNNI node, so it does
not need to be a gateway switch anymore. Modify the default protocol of S5 to
pnni using the following AMI command:
routing domain modify -id <integer> -defaultproto pnni
Reboot switch S5. The state of the network at this point is shown in Figure C.7.
Converting FT-PNNI
and PNNI Networks
PNNI node
PNNI link Fore Area A4
Fore Level L4 S2 S3
FT-PNNI node
S1 A5
FT-PNNI link L5
S4 S5
Figure C.7 - S4 Changed to a Gateway Switch and S5 to PNNI

Reboot switch S2. Upon the reboot, S2 will come up with a FT-PNNI node in area
4 and a PNNI node in area 5. The only FT-PNNI links left in the network are
between S1 and S2 and between S1 and S4.

7. At this point, S3 no longer has any links attached to its FT-PNNI node, so it does
not need to be a gateway switch anymore. Modify the default protocol of S3 to
pnni using the following AMI command:
Reboot switch S3.

8. Since S1 is the last switch in the network to have its default protocol modified, it
can be changed directly to a PNNI switch rather than to a gateway switch first and
then to a PNNI switch. So, modify the default protocol of S1 to pnni using the fol-
lowing AMI command:
Reboot switch S1. Now all of the links have been switched over to PNNI.
9. Convert both S2 and S4 to PNNI by modifying the default protocol of both S2 and
S4 to pnni using the following AMI command:
Reboot switches S2 and S4. The final state of the network upon completion of the
conversion is shown in Figure C.8.
PNNI node
PNNI link S2 S3
Fore Area A5
S1 Fore Level L5
S4 S5
Figure C.8 - A Completely Converted PNNI Network
C - 10 ATM Switch Network Configuration Manual

C.3 Migration of a Hierarchical FT-PNNI Network

This section discusses the migration of a hierarchical FT-PNNI network to a FORE hierarchical
PNNI network. It is assumed that all switches in your network are FORE switches. If your net-
work contains a contiguous backbone, use the instructions found in Section C.3.1. If your net-
work does not contain a contiguous backbone, use the instructions found in Section C.3.2.
C.3.1 Migration of a Hierarchical FT-PNNI Network with a

Contiguous Backbone
There are two different ways to change a FT-PNNI network with a contiguous backbone. If
you want to convert your network starting with the backbone, use the method described in
Section C.3.1.1. If you want to convert your network starting with the peer groups, use the
method described in Section C.3.1.2. You can use either method.
C.3.1.1 Migration Starting with the Backbone

This section gives a detailed example of how to convert a hierarchical FT-PNNI network with
Converting FT-PNNI
and PNNI Networks
a contiguous backbone as shown in Figure C.9 to a PNNI network. In this figure, only the bor-
der nodes are shown. The individual nodes within each peer group are not shown.

FT-PNNI node PG A
FT-PNNI link
B.1 A.1
PG B
C.1
C.2
PG C
Figure C.9 - Hierarchical FT-PNNI Network with 3 Peer Groups and a Contiguous Backbone
ATM Switch Network Configuration Manual C - 11

C.3.1.1.1 Migration Overview

The following basic steps are involved in the migration. Each of these steps is described in
detail in the following sections. It is recommended that you read the entire section before
attempting to change over your network.
2. Convert the backbone of the FT-PNNI network.
3. Convert the individual peer groups of the FT-PNNI network.
C.3.1.1.1.1 Upgrade the Switches

Upgrade all of the switches to ForeThought 6.1.x using the following AMI command:
For more information about upgrading your switches, see Chapter 4 of the ATM Switch
Installation and Maintenance Manual. After the upgrade, there is only one FT-PNNI area with
area ID 4 and level 4. The three FT-PNNI peer groups are contained within this one area.
C.3.1.1.1.2 Convert the Backbone

C.1 will be the first switch to be converted to a gateway between the existing FT-PNNI area
and the PNNI area to be created. Since the backbone is being migrated to PNNI first, the back-
bone PNNI area should have a higher level (numerically smaller) than the FT-PNNI area. The
level of the backbone will be 2.
1. Change the peer group ID of the PNNI node on C.1 to D (so that it has a new peer
group ID which is different from the peer group IDs of the three existing FT-PNNI
peer groups: A, B, and C). Although the PNNI node in C.1 will not be activated
until the default protocol gets changed to gateway, it is possible to change the level
of the PNNI node while it is still down. So, change the level of the PNNI node at
switch C.1 to 2. Use the following AMI commands to change both:
routing pnni node modify -index <integer> -adminstatus down
routing pnni node modify -index <integer> -peergroupid d -forelevel 2
routing pnni node modify -index <integer> -adminstatus up

2. Modify the default protocol of the default domain in C.1 to gateway using the fol-
lowing AMI command:
Reboot switch C.1. Upon the reboot, C.1 will come up with two nodes: one
FT-PNNI node and one PNNI node. There are now two areas: one FT-PNNI area 4
at level 4, and one PNNI area 5 at level 2.
3. A.1 will be the second switch to be converted to a gateway between the existing
FT-PNNI area and the PNNI area to be created. Change the peer group ID of its
PNNI node (currently down) to D, modify the level of its PNNI node to 2, and
modify the default protocol of the default domain in A.1 to gateway using the fol-
lowing AMI commands:
Converting FT-PNNI
and PNNI Networks
Reboot switch A.1. Upon the reboot, the link between the PNNI nodes on C.1 and
A.1 now becomes the first PNNI link in the network. The state of the network at
this stage is shown in Figure C.10.

PNNI node
PG A Area A5
PNNI link Level L2
FT-PNNI peer group boundary A.1
PG D
FT-PNNI node B.1
PG B
FT-PNNI link
C.1
C.2
PG C
Figure C.10 - C.1 and A.1 as Gateway Switches

4. B.1 will be the next switch to be converted to a gateway switch. Administer the
node down. Change the peer group ID of its PNNI node to D, modify the level of
its PNNI node to 2, modify the default protocol of the default domain in B.1 to
gateway, then administer the node up using the following AMI commands:
Reboot switch B.1. Upon the reboot, the link between the A.1 and B.1 becomes a
PNNI link.
5. At this point, the FT-PNNI peer group A is now severed from the rest of the
FT-PNNI area. Since there are now two areas with area ID 4, one of them has to be
renamed. Change the area ID of the FT-PNNI node on A.1 to 2 using the following
AMI command:
routing ftpnni modify -ftpnniforearea 2
Reboot A.1 so the change takes effect. Figure C.11 shows the state of the network
at this point.

PNNI node PG A
PNNI link A.1
A3
FT-PNNI node L4
FT-PNNI link A5
PG B L2
PG D
B.1 C.1
C.2 PG C
Figure C.11 - Peer Group Severed from the Rest of the FT-PNNI Area

6. C.2 is the final backbone switch to be converted to a gateway switch. Change the
peer group ID of its PNNI node (currently down) to D, modify the level of its PNNI
node to 2, and modify the default protocol of the default domain in C.2 to gateway
using the following AMI commands:
Reboot switch C.2. Upon the reboot, all of the backbone links of the network are
converted to PNNI.
7. At this point, the FT-PNNI peer group B is an area by itself. So, the area ID of the
FT-PNNI node on B.1 needs to be changed to 3 using the following AMI command:
routing ftpnni modify -ftpnniforearea 3
Converting FT-PNNI
and PNNI Networks
Reboot switch B.1. The FT-PNNI peer group C is also an area by itself, but it can be
allowed to retain its original area ID of 4. Upon the reboot of switch B.1, the migra-
tion of the backbone is complete.

C.3.1.1.1.3 Convert the Individual Peer Groups

Now each individual peer group needs to be converted to PNNI. In this case, peer group C is
being used as an example. The other peer groups would be converted using the same steps
outlined here. Figure C.12 shows an example configuration of the FT-PNNI nodes within peer
group C that need to be converted.

PNNI node C.1
FT-PNNI peer group boundary C.3
FT-PNNI node
FT-PNNI link
C.2
C.4 C.5
A4
L4
C.6
Figure C.12 - Peer Group C before the Conversion to PNNI
C.6 will be the first switch to be converted to a gateway between the existing FT-PNNI area
and the PNNI area. This area should be at a lower level. The level of this peer group will be 6.
1. Modify the level of the PNNI node (currently down) in C.6 to 2, change the area ID
of its PNNI node to 6, and modify the default protocol of the default domain in C.6
to gateway using the following AMI commands:
routing pnni node modify -index <integer> -forelevel 6 -forearea 6
Reboot switch C.6. Upon the reboot, the first PNNI node is created in area 6. At this
point, this node has no links attached to it.

2. Convert C.5 to a gateway switch by changing the level of its PNNI node to 6,
changing the area ID of C.5 to 6, and modifying the default protocol of the default
domain in C.5 to gateway using the following AMI commands:
Reboot switch C.6. Upon the reboot, the first PNNI link in area 6 comes up between
switches C.6 and C.5.
3. Convert C.4 to a gateway switch by changing the level of its PNNI node to 6,
changing the area ID of C.4 to 6, and modifying the default protocol of the default
domain in C.4 to gateway using the following AMI commands:
Converting FT-PNNI
and PNNI Networks
Reboot switch C.4. Upon the reboot, the link between C.4 and C.6 comes up as
PNNI in area 6. The links between C.4 and C.2 and between C.4 and C.1 will
attempt to come up as PNNI (because these links are between two gateway
switches), but they will not become operational (because the PNNI links will not
reach the two-way-inside hello state).
4. Since C.6 does not need to be a gateway switch anymore, change the default pro-
tocol of the default domain in C.6 to pnni using the following AMI command:
Reboot switch C.6.

5. Since C.3 is the last FT-PNNI switch, convert it directly to a PNNI switch by chang-
ing the level of its PNNI node to 6, changing the area ID of C.3 to 6, and modifying
the default protocol of the default domain in C.3 to pnni using the following AMI
commands:
Reboot switch C.3. Upon the reboot, the link between C.3 and C.1 will attempt to
become PNNI, but will not become operational. At this point, the peer group is
temporarily separated from the backbone and other peer groups in the network as
shown in Figure C.13. It will be restored to full connectivity when C.1 and C.2 have
their second PNNI nodes created.
PNNI node C.1

PNNI link
Inoperational PNNI link C.3
C.5
C.2
C.4 C.6
A6
L6
PG C
Figure C.13 - C.1 and C.2 Not Part of Peer Group C
6. Since C.5 and C.4 do not need to be gateway switches anymore, change the default
protocol of the default domain in C.5 and in C.4 to pnni using the following AMI
command:
Reboot switch C.5 and switch C.4.

7. Change the default protocol of the default domain in C.1 to pnni using the follow-
ing AMI command:
Reboot switch C.1.

8. Create a second PNNI node in C.1 with a node index of 2, the peer group ID set to
C, area 6, and level 6 using the following AMI commands:
routing pnni node new -index 2 -peergroupid c -forelevel 6 -forearea 6
Reboot switch C.1. Then, modify the interfaces on C.1 corresponding to the links
to C.4 and C.3 and attach them to node 2 on C.1. Use the following AMI command
for each link:
routing pnni interface modify -atmif <AtmIf> -vpi <integer>

-nodeindex 2
Converting FT-PNNI
and PNNI Networks
This brings the links to C.3 and C.4 back to being operational.
9. Change the default protocol of the default domain in C.2 to pnni using the follow-
ing AMI command:
Reboot switch C.2.

C, area 6, and level 6 using the following AMI commands:
routing pnni node new -index 2 -peergroupid c -forelevel 6 -forearea 6
Reboot switch C.2.

11. Modify the interfaces on C.2 corresponding to the link to C.4 and attach it to node
2 on C.2. Use the following AMI command for each link:
routing pnni interface modify -atmif <AtmIf> -vpi <integer>

-nodeindex 2
Reboot C.2. This brings the link between C.2 and C.4 back to being operational.
12. Upon reboot of C.1, C.3 and C.4 will no longer need to be gateway switches. So,
modify the default protocol to PNNI using the following AMI command on each
switch:
Reboot both C.3 and C.4.

This completes the migration of peer group C to PNNI. Peer groups A and B can be migrated
to PNNI in a similar way using the steps found in Section C.3.1.1.1.3. Once they are changed
over to PNNI, the entire conversion is complete.
After the migration, the network has four areas at two discrete levels. Each area is a PNNI peer
group. The final state of the network upon completion of the conversion is shown in Figure
C.14.
PNNI node
PNNI peer group boundary PG A
A3
L4
A2
L4 A6
PG D L4
PG B
A5 PG C
L2
Figure C.14 - A Completely Converted PNNI Network

C.3.1.2 Migration Starting with the Peer Groups

This section gives a detailed example of how to migrate a hierarchical FT-PNNI network with
a contiguous backbone as shown in Figure C.9 to a PNNI network. In that figure, only the bor-
der nodes are shown. The individual nodes within each peer group are not shown.
C.3.1.2.1 Overview of the Migration

The following basic steps are involved in the migration. Each of these steps is described in
detail in the following sections. It is recommended that you read the entire section before
attempting to change over your network.
2. Convert the individual peer groups of the FT-PNNI network.
3. Convert the backbone of the FT-PNNI network.
C.3.1.2.1.1 Upgrade the Switches

Upgrade all of the switches from to ForeThought 6.0.x using the following AMI command:
Converting FT-PNNI
and PNNI Networks
For more information about upgrading your switches, see Chapter 4 of the ATM Switch
Installation and Maintenance Manual.
C.3.1.2.1.2 Convert Peer Group C

1. Peer group C is the first to be converted. Start with switch C.6. Modify the default
protocol of the default domain in C.6 to gateway using the following AMI
command:
Reboot the switch. This creates the first PNNI node in the network with a default
area 5 and level 5 as shown in Figure C.15.


PNNI node
PNNI link
FT-PNNI node
PG A
FT-PNNI link
A4 A.1
L4 C.1 C.3 PG C
C.2
PG B
C.5
C.4
B.1
C.6
A5
L5
Figure C.15 - C.6 as a Gateway
lowing AMI command:
Reboot switch C.5. This creates the second PNNI node in the network and the first
PNNI link in area 5 between C.6 and C.5.
lowing AMI command:
Reboot switch C.3. This creates the second PNNI link in the network between C.5
and C.3.

4. At this point, C.5 does not have any more FT-PNNI links left. So, modify the
default protocol of the default domain in C.5 to pnni using the following AMI
command:
Reboot switch C.5.

5. C.4 is the next switch to be converted. Modify the default protocol of the default
domain in C.4 to gateway using the following AMI command:
Reboot switch C.4.

6. Since C.6 has now lost its last FT-PNNI link, change the default protocol of the
default domain in C.6 to pnni using the following AMI command:
Reboot switch C.6.
Converting FT-PNNI
and PNNI Networks
7. C.1 is the next switch to be changed to a gateway. Since the FT-PNNI peer group of
C will become partitioned when C.1 becomes a gateway switch, it is a good idea to
change the peer group ID of the FT-PNNI node on C.1 so that other peer groups
will not attempt to use peer group C as a transit peer group while computing inter-
peer group routes. Change the peer group ID of the FT-PNNI node on C.1 to E
(something other than C) using the following AMI commands:
routing ftpnni modify -peergroupmask <integer>

routing ftpnni modify -switchprefix <Physical Address>
Answer no to the question of whether or not you want to the switch to be rebooted.
8. Modify the default protocol of the default domain in C.1 to gateway as follows:
Now reboot switch C.1. Two more PNNI links between C.1 and C.3 and between
C.1 and C.4 are created.
9. Since C.3 no longer needs to be a gateway switch, modify the default protocol of
the default domain in C.3 to pnni using the following AMI command:
Reboot switch C.3.

10. C.2 is the last switch in the peer group to become a gateway. Modify the default
protocol of the default domain in C.2 to gateway using the following AMI
command:
Upon the reboot of C.2, the last FT-PNNI link (between C.2 and C.4) in the peer
group becomes a PNNI link.
11. Modify the default protocol of the default domain in C.4 to pnni using the follow-
ing AMI command:
Reboot switch C.4.

This completes the migration of peer group C to PNNI. The current state of the network is
shown in Figure C.16.

PNNI node
PNNI link
FT-PNNI node PG A
FT-PNNI link
C.1
A4 A5
L4 A.1 L5
C.3 PG C
PG B
C.5
C.2
B.1
C.4
C.6
Figure C.16 - Peer Group C Fully Migrated to PNNI

C.3.1.2.1.3 Convert Peer Group A

Peer group A is chosen next to be migrated to PNNI. Step step 7 from the conversion of peer
group C is unnecessary in peer group A’s case because there is only one border node in this
peer group, and, therefore, the peer group does not get partitioned during the migration. Also,
ensure that the area ID of the PNNI nodes in peer group A is different from the area ID of peer
group C. Make 6 the area ID of peer group A and make the level 5 (the same level as peer
group C).
Migrate all of the nodes in peer group A using the same method that was used to migrate peer
group C in Section C.3.1.2.1.2.
C.3.1.2.1.4 Convert the Backbone

Once all of the nodes in peer group A have been modified, the link between A.1 and C.1 will
become an inoperational PNNI link because it is between two gateway switches and because
the peer group IDs of the PNNI nodes on A.1 and C.1 do not match. However, connectivity
between the peer groups is still possible through the FT-PNNI backbone.
1. Modify the default protocol of the default domain in C.1 to pnni using the follow-
ing AMI command:
Converting FT-PNNI
and PNNI Networks
Reboot switch C.1.

D, area 3, and level 2 using the following AMI commands:
routing pnni node modify 2 -peergroupid d -forelevel 2 -forearea 3
Reboot switch C.1. This is the first node in the backbone area.
3. Modify the interface on C.1 corresponding to the link to A.1 and attach it to the
newly-created node 2 on C.1 using the following AMI command:
conf atmroute pnni interface modify -atmif <AtmIf> -vpi <integer>

-nodeix 2

4. Change the default protocol of the default domain in A.1 to pnni using the follow-
ing AMI command:
Reboot switch A.1. The FT-PNNI link between A.1 and B.1 is lost at this stage and
connectivity between peer group A and the rest of the network is severed.
5. Create a second PNNI node in A.1 with a node index of 2, the peer group ID set to
D, area 3, and level 2 using the following AMI commands:
Reboot switch A.1.

6. Modify the interface on A.1 corresponding to the link to C.1 and attach it to the
newly-created node 2 on A.1 using the following AMI command:
conf atmroute pnni interface modify -atmif <AtmIf> -vpi <integer>

-nodeix 2
At this point, the connectivity between peer groups A and C should be restored,
but peer groups A and B still remain disconnected. The current state of the network
is shown in Figure C.17.


PNNI node
PNNI link
FT-PNNI node
PG A
FT-PNNI link
A6 A.1
L5
C.1
A3 A5
L2 L5 C.3 PG C
PG B
C.5
A4 C.2
B.1
Converting FT-PNNI
and PNNI Networks
L4
C.4
C.6
Figure C.17 - Peer Group B Disconnected from Peer Group A
To quickly restore the connectivity, the rest of the backbone should be migrated to PNNI.
7. Change the peer group ID to D, the level to 2, and area to 3 on the PNNI node in
B.1, and change the default protocol of the default domain in B.1 to gateway using
the following AMI commands:
Reboot switch B.1. The link between A.1 and B.1 should now be restored as a PNNI
link and connectivity resumes between A and B.

8. C.2 no longer needs to be a gateway switch. Modify the default protocol of the
default domain in C.2 to pnni using the following AMI command:
Reboot switch C.2. Upon the reboot, the backbone and peer groups A and C are
now running PNNI and only peer group B needs to be migrated. The current state
of the network is shown in Figure C.18.

PNNI node
PNNI link
FT-PNNI node
PG A
FT-PNNI link A6
L5
A.1
PG B
A5
A4 C.1 L5
L4 B.1 C.3 PG C
PG D
A3
L2
C.2 C.5
C.4
C.6
Figure C.18 - A Migrated Backbone
C.3.1.2.1.5 Convert Peer Group B

Migrate all of the nodes in peer group B using the same method that was used to migrate peer
group C in Section C.3.1.2.1.2. Once peer group B has been converted, the migration to PNNI
is complete.

C.3.2 Migration of a Hierarchical FT-PNNI Network with a Non-

Contiguous Backbone
This section gives a detailed example of how to migrate a hierarchical FT-PNNI network with
a non-contiguous backbone as shown in Figure C.19 to a PNNI network. In this figure, only
the border nodes are shown. The individual nodes within each peer group are not shown.

FT-PNNI node
FT-PNNI link
PG A
A.2
A.1
B.1 A4
L4
C.1
Converting FT-PNNI
and PNNI Networks
PG B
B.2
PG C
C.2
Figure C.19 - Hierarchical FT-PNNI Network with 3 Peer Groups and a Non-contiguous Backbone
Upon the complete migration of this network to PNNI, the network is modified as follows:
• Make peer group A the backbone area in a split-switch based hierarchical net-
work.
• Peer group A would become a PNNI area with an area ID of 2 and level 2 with a
single PNNI peer group in the new network.
• Peer group B would become a PNNI area with an area ID 1 and level 4 (a lower
level than 2) with a single PNNI peer group.
• Peer group C would become a PNNI area with an area ID 3 and level 4 (a lower
level than 2) with a single PNNI peer group.

• The link between B.2 and C.2 in the FT-PNNI network can be placed in a new area
(with peer group ID D), which is a higher level area connecting peer groups B and
C and acts as a back door entry between the two peer groups. This provides the
redundancy that this link offered in the original FT-PNNI network.
Figure C.20 shows the state of the network on the completion of the migration.

PNNI node
PNNI link A.1 A.2
PG A
A2
L2 A.4
B.1 C.1
A4
L2
PG B B.2
A1 C.2 PG C
L4
D.2 D.1 A3
L4
PG D
Figure C.20 - Hierarchical PNNI Network after Migration

C.4 Migration of a Non-hierarchical PNNI Network to a

Lop-sided Hierarchical PNNI Network
This section discusses the conversion of a non-hierarchical (single peer group) PNNI network
to a lop-sided hierarchical PNNI network. A lop-sided hierarchical network is one in which
lowest level nodes and Logical Group Nodes (LGNs) are acting as peers in a single peer
group. In this example, all switches in the network are FORE switches.
C.4.1 Migration Overview

This section gives a detailed example of how to change the non-hierarchical PNNI network
shown in Figure C.21 to a lop-sided hierarchical PNNI network.
2. Select a peer group leader (PGL) candidate. Set its node PGL priority to a non-zero
value.
3. Configure a parent for that node.
Converting FT-PNNI
and PNNI Networks
4. Verify the election on all of the switches.
5. Change the interface on the node that resides in the higher level peer group.
6. Verify that this interface becomes an outside link.
7. Administer the de-activated node down.
8. Verify that the LGN has set up an SVCC RCC.
9. Repeat step 1 through step 8 for each peer group.
PMP connections are not supported on an

NOTE ASX-1000, ASX-1200, or TNX-1100 that is split
into multiple peer groups. See Section C.1 for
more information.

C.4.2 Detailed Migration Example

This section gives a detailed example of how to change the non-hierarchical PNNI network
shown in Figure C.21 to a lop-sided hierarchical PNNI network.
This migration procedure works only when

NOTE addresses and address prefixes in the lower level
peer groups (B.X and B.Y in Figure C.21) can be
summarized by the higher level peer group (B in
Figure C.21). If they are not summarized by the
higher level peer group (e.g., if B.X was peer
group F, B.Y was peer group G, and B was peer
group H), the procedure can still be used, but an
extra step is needed after step 5. The PG IDs of
B.3 and B.2 in the higher level peer group (B)
needs to be modified so that LGN B.X can reach
those addresses.

PNNI node B.4
PNNI link B.3 B.5
PG B
B.1 Level 80
B.6
B.2
B.X.2 B.Y.3
Level 80
B.X.1
Level 80 B.Y.1 B.Y.2
PG B.X B.X.3 PG B.Y
Figure C.21 - Non-hierarchical PNNI Example
1. Upgrade each switch in the network to ForeThought 6.1.x as follows:
For more information about upgrading your switches, see Chapter 4 of the ATM
Switch Installation and Maintenance Manual.

2. Start by migrating peer group B.X. Select a peer group leader (PGL) candidate and
set its node PGL priority to a non-zero value so that it gets elected as PGL. In this
example, B.X.1 is used.
routing pnni node modify 1 -pglprio 5
3. Configure a parent for that node using the following command:
routing pnni node newparent -parent <integer> -child <integer>

-level <level>
routing pnni node newparent -parent 2 -child 1 -level 72
4. Verify the election on all of the switches in peer group B.X using the following
command:
routing pnni node show -pgl
On the node that has been elected PGL, the PglState should say operPgl. At
Converting FT-PNNI
this point, an LGN B.X representing peer group B.X has been activated by the PGL
and PNNI Networks

B.X.1 as shown in Figure C.22.

PNNI node B.4
B.3 B.5
PNNI link
PG B
PGL
B.1 Level 72
B.6
B.2
B.X.2 B.Y.3
Level 80
B.X.1
B.Y.1 B.Y.2
Level 80
B.X.3
PG B.X PG B.Y
Figure C.22 - B.X.1 Elected as PGL

5. On the split switch containing B.X.2 (lower level node) and B.1 (higher level node),
move any links attached to the higher level node (B.1) to the lower level node
(B.X.2) as shown in Figure C.23. Change the interfaces on node 1 (B.X.2) using the
following commands:
routing pnni interface modify -atmif 1C1 -vpi 0 -nodeindex 1
routing pnni interface modify -atmif 1D1 -vpi 0 -nodeindex 1

PNNI node B.4
B.3 B.5
PNNI link
PG B
PGL
B.1 1C1 0 Level 72
Migrated PNNI link B.6
1D1 0
B.2
B.X.2 B.Y.3
Level 80
B.X.1
B.Y.1 B.Y.2
Level 80 B.X.3
PG B.X PG B.Y
Figure C.23 - Migrating the Links
6. These links become PNNI outside links because they connect nodes in two differ-
ent peer groups (PG B.X and PG B). Verify that 1C1 0 and 1D1 0 become outside
links by noting that the HelloSt for each says twoWayOutside using the follow-
ing command:
routing show
7. Node B.1 (node 2 on the split switch) can now be de-activated. Administer this
node down as follows:
routing pnni node -index 2 -adminstatus down
The peer groups now look like the example in Figure C.24.


PNNI node B.4
B.3 B.5
PNNI link
PG B
PGL Level 72
B.X B.6
LGN
B.2
B.Y.3
B.X.2 Level 80
B.Y.1 B.Y.2
B.X.1
Level 80 PG B.Y
B.X.3
PG B.X
Figure C.24 - Node B.1 is De-activated and LGN B.X Represents Peer Group B.X
Converting FT-PNNI
and PNNI Networks
8. LGN B.X (residing at the same switch as B.X.1) now sets up SVCC RCCs to B.2 and
B.3 and initiates a peer relationship with them. LGN B.X now becomes an integral
part of peer group B. Verify that the LGN has set up an SVCC RCC with B.2 and
B.3 as follows:
a. On the LGN switch, use the following command:
routing pnni svccrcc
The HelloState should say twoWayInside because the switches are within the
same peer group.
b. On switches B.2 and B.3, use the following command:
routing pnni svccrcc
The HelloState should say twoWayInside because the switches are within the
same peer group.
9. Repeat step 1 through step 8 on peer group B.Y to convert it in a similar manner so
that B.Y.3 is the PGL in peer group B.Y and so that LGN B.Y is created as shown in
Figure C.25.


PNNI node B.4
B.3 B.5
PNNI link
PG B
PGL Level 72
B.X B.Y
LGN
B.2
B.X.2 B.Y.3
Level 80
B.X.1
B.Y.1 B.Y.2
Level 80
B.X.3
PG B.X PG B.Y
Figure C.25 - Completed Conversion
The migration is now complete.

APPENDIX D Configuring FramePlus
Modules
FramePlus™ network modules (NMFR-4/DS1A and NMFR-4/E1A) are four-port interfaces

that support interworking between ATM and Frame Relay, or ATM and Frame-based UNI
(FUNI).
All four ports of the DS1 FramePlus Network Module can support fractional DS1 services (n x
64 Kbps) where 1 to 24 contiguous or non-contiguous DS0 channels are mapped to a single
ATM port.
All four ports of the E1 FramePlus Network Module can support fractional E1 services (n x 64
Kbps) where 1 to 31 contiguous or non-contiguous DS0 channels are mapped to a single ATM
port. E1 also supports an unframed mode in which all 32 channels are mapped to a single
ATM port.
This appendix provides an overview of the FramePlus network module and provides configu-
ration examples. Topics that are covered include the following:
• Section D.1 - Frame Relay Overview
• Section D.2 - Configuring the Module
• Section D.3 - Profiles and Services
• Section D.4 - Configuring Frame Relay
• Section D.5 - Configuring FUNI
• Section D.6 - Upgrading the FramePlus Network Module Software
FramePlus Modules
Configuring
ATM Switch Network Configuration Manual D-1

Configuring FramePlus Modules
D.1 Frame Relay Overview

Serving as an interface between user and network equipment, Frame Relay provides a means
for statistically multiplexing many logical data channels (or virtual circuits) over a single
physical transmission link. Frame Relay is ideal for supporting multiple data streams because
it provides flexible and efficient utilization of available bandwidth.
D.1.1 Interworking Function (IWF)

When connecting an ATM Customer Premise Equipment (CPE) and a Frame Relay CPE, each
device has no knowledge that the distant device supports a different protocol. The FramePlus
network module performs a service interworking function (IWF) to terminate the respective
Frame Relay and ATM services and to perform appropriate interworking conversions before
transmitting the CPE information. IWF uses two different modes of operation to address
interoperability between upper layer user protocol encapsulation techniques on the FramePlus
network module: translation mode and transparent mode.
D.1.1.1 Translation Mode

When the two CPEs support different standard encapsulation protocols, a translation, or re-
encapsulation, must take place in the CPE payload portion of the cells that are being transmit-
ted. Because Frame Relay supports RFC 1490 encapsulation and ATM supports RFC 1483
encapsulation, the information in the frames is converted to cells and vice versa.
D.1.1.2 Transparent Mode

When both CPEs support the same encapsulation protocol, there is no need for any translation
or re-encapsulation of the CPE payload portion of the cells that are being transmitted. There-
fore, the traffic is sent “transparently.” Transparent mode must be used if the CPEs support
non-standard encapsulation protocols. When the transparent method of IWF is used, there is a
mapping that occurs of each Frame Relay DLCI to an ATM VPI, VCI combination.
One example would be if proprietary encapsulation techniques are being used. A second
example would be if there is FR/CPE equipment at both edges of the network with ATM pro-
viding the transport, there would be no need for any translation.
D-2 ATM Switch Network Configuration Manual

D.2 Configuring the Module

Before configuring any services or PVCs, you should divide the buffer space between the high
and low priority buffers as described in Section D.2.1. Cells are read from the high priority
buffer first. After the high priority buffer is emptied, cells are read from the low priority
buffer. Then you should set the EPD/PPD thresholds for the high and low priority buffers as
described in Section D.2.2.
D.2.1 Dividing the Buffer Space

The FramePlus network module has two priority buffers: high and low. The total, combined
buffering space available is 32,768 cells. The amount of buffering space is split between the
two buffers using hardware netmod fram modify -epdppdbufferproportion. There
are four fixed configuration models for partitioning the buffer sizes as shown in Table D.1. The
default model is highzero.
Table D.1 - Buffer Models
Model High Priority Buffer Low Priority Buffer

highzero 0 cells 32,768 cells
high1quarter 8,192 cells 24,576 cells
If you choose highzero, you want all of the buffering space to go in the low priority buffer and
none in the high priority buffer. If you choose high1quarter, then one fourth of the buffering
FramePlus Modules
space goes in the high priority buffer and three fourths goes in the low priority buffer, etc.
Configuring
Before you can change the buffer allocation, you must first take the network module out of
service by administering it down as follows:
myswitch:hardware netmod-> modify -module 1d -admin down
Disabling the network module will destroy all
existing connections going through it.
Disable the network module [n]? y

myswitch:hardware netmod-> fram
myswitch:hardware netmod fram-> modify -module 1d -epdppdbufferproportion
high2quarter
You need to administer the network module back up after you make the change.
myswitch:hardware netmod-> modify -module 1d -admin up

D.2.2 Setting the Thresholds

After the buffer space is configured, set the various discard thresholds under hardware
netmod fram modify.
It is important that you follow the order shown

NOTE in this section for configuring the thresholds. The
calculation of each threshold is dependent on the
previous one.
For this buffer configuration example, assume you have set the buffers to use high2quarter,
meaning that both the high and low priority buffers have 16, 384 cells as shown in Figure D.1.
16,384 cells 16,384 cells
High Priority Buffer Low Priority Buffer
Figure D.1 - Buffer Sizes Configured Using the Hardware Netmod Fram Modify Command
The number of cells shown in this section for

NOTE buffer sizes are approximations.

D.2.2.1 Noting the CLP0PPD Threshold

Each buffer has four different thresholds. The first is for Partial Packet Discard (PPD) on cells
with CLP=0. This threshold is automatically set for you as 87.5% of the size of each buffer and
cannot be changed. (It is not displayed in AMI currently, but you need to know this value
because the other thresholds are calculated on the remaining buffer size.)
Therefore, in this example, since the total buffer size you set is 16,384 cells then the threshold
for CLP0PPD is 87.5% of each buffer as shown in Figure D.2:
16,384 cells * 0.875 percent = 14,336 cells
14,336 CLP0PPD (fixed - 87.5%) 14,336 CLP0PPD (fixed - 87.5%)
FramePlus Modules
Figure D.2 - CLP0PPD Automatically Calculated
Configuring

D.2.2.2 Configuring the CLP1EPD Threshold

Next, calculate the CLP1EPD threshold for each buffer. You have four choices: 25, 37, 50, or 62
percent. For example, you could set the threshold to 50 percent of the remaining buffer size of
each buffer, which is 14,336, as shown by the arrow in Figure D.3:
14,336 cells * 0.5 = 7,168 cells
7,168 CLP1EPD (50%) 7,168 CLP1EPD (50%)
Figure D.3 - Calculating CLP1EPD

D.2.2.3 Configuring the CLP0EPD Threshold

Now, calculate the CLP0EPD threshold for each buffer. You have four choices: 50, 62, 75, or 87
percent. For example, you could set the threshold to 50 percent of the remaining buffer size of
each buffer, which is 7,168, as shown by the arrow in Figure D.4. Then, add the base of 7,168
back in.
7,168 cells * 0.5 = 3,584 cells + 7,168 cells = 10,752 cells
10,752 CLP0EPD (50%) 10,752 CLP0EPD (50%)
7,168 CLP1EPD (50%) 7,168 CLP1EPD (50%)
Figure D.4 - Calculating CLP0EPD
FramePlus Modules
Configuring

D.2.2.4 Configuring the CLP1PPD Threshold

Finally, calculate the CLP1PPD threshold for each buffer. You have four choices: 50, 62, 75, or
87 percent. For example, you could set it to 50 percent of the remaining buffer size, which is
3,584, as shown by the arrow in Figure D.5. Then, add the base of 7,168 back in as shown:
10,752 cells - 7,168 cells = 3,584 cells
3,584 cells * 0.5 = 1,792 cells
1,792 cells + 7,168 cells = 8,960 cells
10,752 CLP0EPD (50%) 10,752 CLP0EPD (50%)

8,960 CLP1PPD (50%) 8,960 CLP1PPD (50%)
7,168 CLP1EPD (50%) 7,168 CLP1EPD (50%)
Figure D.5 - Calculating CLP1PPD

D.3 Profiles and Services

A profile lets you define a set of information that can be applied to a particular service or PVC
using a single index number, much in the same way that you can define a set of traffic manage-
ment information in a UPC contract and then apply it to a PVC using a single index number.
For most profile types, several default profiles are provided by the switch for your conve-
nience. These default profiles reflect what typical networks may need. You cannot delete the
default profiles. If a service or connection is created without specifying any profile, the first
default profile in the list (index 0) is used.
You can also create new profiles and delete existing profiles, but you cannot modify existing
ones. Each profile has an associated reference count, which identifies the number of services or
connections that are currently using that profile. You cannot delete any profile that is currently
in use (has a non-zero reference count in the Ref field under the profile’s show command).
There are six different profiles that are supported by the FramePlus network module: FRF.8,
EPD/PPD, FRRATE, LMI, service, and FUNI. The FRF.8, and FRRATE profiles can only be
applied to Frame Relay PVCs. The FUNI profiles can only be applied to FUNI PVCs. The
EPD/PPD profiles can be applied to either Frame Relay or FUNI PVCs. The LMI profiles can
only be applied to Frame Relay services. The service profiles can be applied to either Frame
Relay or FUNI services. All profiles can be configured under the services profile menu.
Please see Part 3 of the AMI Configuration Commands Reference Manual for detailed information
about the AMI parameters. Each of these profiles is discussed in the following sections.
D.3.1 EPD/PPD Profile

The Early Packet Discard/Partial Packet Discard (EPD/PPD) profile lets you determine how
selective cell discard is performed. The profile then can be applied on a per-VC basis to a
Frame Relay PVC using the -epdppd <integer> option under services fratm pvc new
FramePlus Modules
or to a FUNI PVC using the -epdppd <integer> option under services funi pvc new.
Configuring
The EPD/PPD profile lets you specify which priority buffer (high or low) will handle the traf-
fic of the given PVC. The different thresholds can be enabled or disabled per PVC. These
thresholds can be configured under hardware netmod fram modify.
D.3.2 FRF.8 Profile

FRF.8 is the Frame Relay Forum’s FR/ATM PVC Service Interworking Implementation Agree-
ment. The FRF.8 profile allows you to define the interworking mappings between Frame Relay
and ATM. The profile then can be applied on a per-VC basis to a Frame Relay PVC using the
-frf8 <integer> option under services fratm pvc new.
The FRF.8 profile lets you decide if the Discard Eligibility (DE) field in every frame is mapped
to the Cell Loss Priority (CLP) field in every ATM cell. It lets you define if the Forward Explicit
Congestion Notification (FECN) field in every frame is mapped to the EFCI field of every ATM
cell. It also lets you define the possible encapsulation translation protocols that can be used.

D.3.3 Frame Relay Rate Profile

The Frame Relay rate profile allows you to manage and define rate enforcement characteris-
tics that can be applied on a per-VC basis to a Frame Relay PVC using the -frrate
<integer> option under services fratm pvc new.
You can explicitly enable rate enforcement in both the ingress (Frame Relay to ATM) and the
egress (ATM to Frame Relay) direction on a FramePlus network module. Ingress rate enforce-
ment is configured using the parameters inbc, inbe, and incir. Egress rate enforcement is
configured using the parameters outbc, outbe, and outcir. If no rate enforcement parame-
ters are specified, the default values are used.
The egress rate enforcement also provides an option by which frames on a per-VC basis can be
buffered as opposed to being discarded when thresholds have been crossed. This buffering
option can increase goodput. The selection of buffering is configured at a service level, but
only takes effect on a connection if egress rate enforcement has been enabled. The amount of
buffering assigned to a connection is finite and is determined internally by the system. If the
buffering has been selected, then the system holds traffic in excess of Bc+Be for a period of Tc
(= Bc/CIR), before trying to resend the buffered traffic. If the buffer becomes full, excess traffic
is discarded. The rate adaptation mechanism is activated for a PVC after an ingress frame,
addressed to the PVC, has been received with BECN set.
The Frame Relay rate profile performs the exact same function that UPC contracts do for
ATM. If you want to use the overbooking option in the service profile, then you need to apply
a Frame Relay rate profile. Bc is the committed burst, Be is the excess burst, and CIR is the
committed information rate. CIR is derived as follows:
CIR = Bc / time period
For example, if the Bc is 10 bits and the time period is 125 microseconds, then the CIR is 10 x
8,000 or 80,000 bps.
EIR is the excess information rate and is derived as follows:
EIR = Be / time period
For example, if the Be is 10 bits and the time period is 125 microseconds, then the EIR is 10 x
8,000 or 80,000 bps.
CIR is very similar to VBR SCR and EIR is very similar to VBR PCR.
D - 10 ATM Switch Network Configuration Manual

D.3.4 LMI Profile

The Link Management Interface (LMI) profile lets you define the version of LMI that is sup-
ported by a particular Frame Relay service. The profile can be applied on a per-service basis
using the -lmi <integer> option under services fratm new. If you do not want to use
LMI, then enter the none for the -flavour <flavour> option under services profile
lmi new and use this profile when creating your services.
LMI serves a function that is similar to ILMI for ATM. LMI signalling occurs on DLCI 1023.
LMI timers poll the network and report if services are up or down. This information can be
optionally translated into ATM OAM F5 cells so that when a DLCI goes down, it triggers an F5
cell for the corresponding PVC. This profile lets you specify the values of the various LMI tim-
ers. Please see Chapter 2 in Part 3 of the AMI Configuration Commands Reference Manual for the
default values and the ranges for the timers.
D.3.5 Service Profile

The generic service profile lets you specify service attributes that are common to Frame Relay
interworking services and FUNI services. This profile can be applied on a per-service basis to
a Frame Relay service using the -service <integer> option under services fratm new
or to a FUNI service using the -service <integer> option under services funi new.
This profile is important when using multiple timeslots on a port because you must ensure
that the access rate of the referenced service profile index is large enough to support them. For
example, the default access rate is 64 Kbps. If you leave the rate at 64, you must only use a sin-
gle timeslot. If you set the rate to 128 Kbps, you must use two timeslots. If you set the rate 1536
Kbps, you must use 24 timeslots, etc.
This profile also lets you define the maximum frame size, the maximum number of DLCIs/
PVCs supported, and input and output bandwidth overbooking for a given service.
FramePlus Modules
Configuring
D.3.6 FUNI Profile
The FUNI profile lets you determine the VPI, VCI range to use for FUNI connections. The pro-
file then can be applied on a per-service basis to a FUNI service using the -funi <integer>
option under services funi new.
D.3.7 Services
A service is a grouping of timeslots on a port. In this respect, a service is similar to an ATM
PVP. Multiple DLCIs (connections) can exist within each service.
ATM Switch Network Configuration Manual D - 11

D.4 Configuring Frame Relay

To configure Frame Relay on a FramePlus network module, these steps must be performed in
the following order:
1. Choose or create any profiles that you wish to use.
2. Create each service that you need and associate the profile(s) with it.
3. Create each PVC and associate the profile(s) and service(s) with it.
These steps are described in detail in the following sections. There are also steps described for
configuring SPANS SPVCs and PNNI SPVCs on a FramePlus network module.
A FramePlus network module can run either a

NOTE Frame Relay over ATM or Frame-based User to
Network Interface (FUNI) application. Because
Frame Relay is the default, it is assumed that you
do not need to change the application from FUNI
to Frame Relay at this point. If you need to
change the application, see the instructions in
Section D.5.1. The Appln field under hardware
netmod application show displays which
application is currently running (either
FRAM_FR_ATM_APPL for Frame Relay or
FRAM_FUNI_ATM_APPL for FUNI).
D.4.1 Choosing Frame Relay Profiles

First, you need to decide which profiles you want to use. For most profile types, several
default profiles are provided for your convenience. If a service or connection is created with-
out specifying any profile, the first default profile in the list (index 0) is used.
To display the default profiles, use the show command under each profile; e.g., services
profile epdppd show. If these defaults do not match your network’s needs, you can create
your own profile using the new command under each profile; e.g., services profile
epdppd new. Create the new profiles for each profile type that you are going to use. Once you
have determined which default profiles to use, you can move on to Section D.4.2.

D.4.2 Creating Frame Relay Services

Create any Frame Relay services that you want and apply the LMI and service profiles. For
example:
The collection of statistics is disabled, by default,

NOTE at the service level. It is enabled, by default, at
the module level. If you want to collect statistics,
you must enable that functionality here at the
service level as well.
myswitch:services fratm-> new

Usage:
[[-id] <text>] Service ID (default: “”)
[-timeslotstr] <text> Timeslots
[[-lmi] <integer>] LMI Profile Index (default: 0)
[[-service] <integer>] Service Profile Index (default: 0)
[[-egress_re] (enabled|disabled)] Egress Re (default: disabled)
[[-state] <ifadminstatus>] Admin state (default: up)
[[-traps] (enabled|disabled)] Traps (default: enabled)
[[-stats] (enabled|disabled)] Stats (default: disabled)
myswitch:services fratm-> new -iatmif 4c1 -timeslotstr 1 -name service_a -stats

enabled
The newly created service id is 4C1:00
FramePlus Modules
Configuring
myswitch:services fratm-> new -iatmif 4c1 -timeslotstr 2 -lmi 1 -name service_b
-stats enabled
myswitch:services fratm-> new -iatmif 4c1 -timeslotstr 3 -name service_c -stats

enabled
myswitch:services fratm-> show

SvcId SvcInd TimeSlot Admin EgressRE Lmi Ser Traps Stats Name
4C1:00 1066 1 up disabled 0 0 enabled enabled service_a
4C1:01 1068 2 up disabled 1 0 enabled enabled service_b
4C1:02 1070 3 up disabled 0 0 enabled enabled service_c

D.4.2.1 Creating an Unstructured E1 Service

If you want to create an unstructured E1 service, perform the following steps:
1. Turn off CRC4 for the port using the following command:
interfaces dsx1 modify -hwportname <BNP> -linetype e1
2. Create a service profile with an access rate of 2048 Kbps as follows:
services profile service new -index <integer> -accrate 2048
3. Create a new service using the port from step 1 as the -iatmif, specifying the
timeslots (-timeslotstr) as either all or 0-31, and using the service profile
from step 2 for -service, as follows:
services fratm new -iatmif <AtmIf> -timeslotstr <text> -service

<integer>

D.4.3 Creating Frame Relay PVCs

Now that the profiles and services have been created, you can create your PVCs and apply the
EPD/PPD, FRF.8, and FRRATE profiles to the PVCs. For example:
myswitch:services fratm pvc-> new

Usage:
[-id] <text> Service ID
[-dlci] <integer> DLCI
[[-oatmif] <AtmIf>] Ingress Output AtmIf
[[-ovpi] <integer>] Ingress Output VPI
[[-ovci] <integer>] Ingress Output VCI
[[-oserviceid] <text>] Output Service ID
[[-odlci] <integer>] Output DLCI
[[-faupc] <UPC Index>] FAUPC (default: 0)
[[-afupc] <UPC Index>] AFUPC (default: 0)
[[-epdppd] <integer>] EPDPPD (default: 0)
[[-state] <ifadminstatus>] Admin State (default: up)
[[-ostate] <ifadminstatus>] Output Admin State (default: up)
[[-frrate] <integer>] FR Rate (default: 0)
[[-frf8] <integer>] FRF8 (default: 0)
[[-ingress_re] (disabled|enabled)] Ingress Re (default: disabled)
To create a regular Frame connection (one end on

NOTE the FramePlus network module and one end on
another network module), the -oatmif, -ovpi,
FramePlus Modules
and -ovci must be specified.
Configuring
To create a port-to-port Frame connection (both
ends on the same FramePlus network module),
the -oserviceid and -odlci must be
specified.
If neither set of parameters is specified, a

dangling PVC is created.
myswitch:services fratm pvc-> new -id 4C1:00 -dlci 100 -oatmif 4d1 -ovpi 0 -ovci
100 -epdppd 1 -name pvc_a
101 -name pvc_b

102 -name pvc_c
myswitch:services fratm pvc-> show

Input Output
SVCId DLCI AtmIf VPI VCI AtmIf VPI VCI Admin Name
4C1:00 100 4C1 0 32 4D1 0 1 up pvc_a
4C1:00 100 4D1 0 1 4C1 0 32 up pvc_a
4C1:01 101 4C1 16 32 4D1 0 100 up pvc_b
4C1:01 101 4D1 0 100 4C1 16 32 up pvc_b
4C1:02 102 4C1 32 32 4D1 0 102 up pvc_c
4C1:02 102 4D1 0 102 4C1 32 32 up pvc_c
To propagate OAM error information from the

NOTE ATM side to the Frame Relay side you need to
enable three parameters:
• Enable -oamf5supervision under

hardware netmod fram modify (this is
enabled by default).
• Enable -aisrdi under interfaces atmif
modify (this is not enabled by default).
• Enable -atmlayeroam under system
modify (this is not enabled by default).

D.4.4 Configuring Frame Relay SPVCs

You can also configure SPVCs (Smart Permanent Virtual Circuits) on a FramePlus network
module. An SPVC is a connection that spans multiple switch fabrics and looks like a PVC at
the local and remote endpoints with an SVC in the middle. If a link carrying an SPVC goes
down and there is an alternate route, then the end switch fabrics of the SPVC automatically
reroute the SPVC around the failed link.
However, the procedure for creating both SPANS SPVCs and PNNI SPVCs on a FramePlus net-
work module is slightly different than it is for other network modules. Section D.4.4.1
describes how to create a Frame Relay SPANS SPVC. Section D.4.4.2 describes how to create a
Frame Relay PNNI SPVC.
D.4.4.1 Creating a Frame Relay SPANS SPVC

To create a Frame Relay SPANS SPVC, perform the following steps:
3. Create a service on the source switch (shark) if you have not already done so as
follows:
shark:services fratm-> new -id 1a1:00 -iatmif 1a1 -timeslotstr 1

NOTICE: SNMP_TRAP: Specific 4 (linkUp) [1037] [2] [1] [Xmit]
The newly created service ID is 1A1:00.
FramePlus Modules
Configuring
4. Create a dangling PVC on the source switch (i.e., do not specify the -oatmif,
-ovpi, and -ovci parameters) as follows:
shark:services fratm-> pvc

shark:services fratm pvc-> new -id 1a1:00 -dlci 100
WARNING: None of output port id, vpi and vci is specified. A dangling
connection will be created.
5. Create a service on the destination switch (tuna) if you have not already done so as
follows:
tuna:services fratm-> new -id 1a1:00 -iatmif 1a1 -timeslotstr 1

6. Create a dangling PVC on the destination switch as follows:
tuna:services fratm-> pvc

tuna:services fratm pvc-> new -id 1a1:00 -dlci 100



SPANS Address from the source switch (shark) as the -srcswitchaddr value as
follows:

-srcswitchaddr 00000038.f21a333b -outatmif 1a1 -outvpi 0 -outvci 32

field as follows:

Local Remote State
1 1A1 0 32 0 1 up

Local Remote State
1 1A1 0 32 0 1 up


-srcswitchaddr 00000038.f21a3656 -outatmif 1a1 -outvpi 0 -outvci 32
FramePlus Modules
Local Remote State
Configuring
1 1A1 0 32 0 1 up

Local Remote State
1 1A1 0 32 0 1 up

D.4.4.2 Creating a Frame Relay PNNI SPVC

To create a Frame Relay PNNI PP SPVC, perform the following steps:
follows:
shark:services fratm-> new -id 1a1:00 -iatmif 1a1 -timeslotstr 1

shark:services fratm-> pvc

shark:services fratm pvc-> new -id 1a1:00 -dlci 100
follows:
tuna:services fratm-> new -id 1a1:00 -iatmif 1a1 -timeslotstr 1

tuna:services fratm-> pvc

tuna:services fratm pvc-> new -id 1a1:00 -dlci 100
6. Find the destination NSAP of the destination ATM interface (1A1 in this example)
on the destination switch as follows:


follows:

field as follows:

-----------------------------------------------------------------------------
1 1A1 0 32 0 require 5 up
0 0 32 0
complete.
FramePlus Modules
Configuring

D.5 Configuring FUNI

To configure FUNI on a FramePlus network module, perform these steps in the following order:
1. Change the application key, if necessary.
2. Choose or create any profiles that you wish to use.
3. Create each service that you need and associate the profile(s) with it.
4. Create each PVC and associate the profile(s) and service(s) with it.
These steps are described in detail in the following sections. There are also steps described for
configuring PNNI SPVCs on a FramePlus network module.
D.5.1 Changing the Application Key

This command lets you configure a FramePlus network module to run either Frame Relay over
ATM or to run frame-based User to Network Interface (FUNI). The network module can only
run one application or the other at a time. However, you can have some network modules run-
ning FUNI and some running Frame Relay in the same switch fabric.
FramePlus network modules run Frame Relay by default. To run FUNI services on the network
module, please contact FORE’s Technical Assistance Center for a valid FUNI key. To change
from FUNI back to Frame Relay, the key is fratm170358. The Application field under
hardware netmod application show displays which application is currently running.
myswitch:hardware netmod application-> show
Netmod Application LED State Status
2B FRAM_FR_ATM_APPL off up Netmod Up & Running
The application key should be specified only if you want to reconfigure the network module
to run a different type of application. When you change the application, the switch deletes all
existing services and PVCs that use a different application, and removes them from the CDB
(i.e., if you are changing from Frame Relay to FUNI, the switch deletes existing Frame Relay
information, and vice versa). To change the application, first admin the module down:
myswitch:hardware netmod-> modify 2B down
WARNING: Disabling the network module will destroy all existing connections
going through it.
Do you wish to continue (y or n)? y
myswitch:hardware netmod-> fram modify 2B -applkey 4a XXXXXXXXXX

INFO: FRAMNM - CDB of slot 2B removed
myswitch:hardware netmod application-> show
Netmod Application LED State Status
2B FRAM_FUNI_APPL off up Netmod Up & Running

D.5.2 Creating the Profiles for FUNI

First, you need to decide which profiles you want to use. A default profile is provided for your
convenience. If a service or connection is created without specifying any profile, the default
profile in the list is used.
To display the default profile, use the services profile funi show command. If this
default does not match your network’s needs, you can create your own using the new com-
mand under each profile; e.g., services funi new. The values that you choose for [-minVci
<integer>] and [-maxVci <integer>] determine the range of VCIs that you can use for
-fvci <integer> under services funi pvc new. (See Part 3 of the AMI Configuration
Commands Reference Manual for more information about the parameters.) Once you have deter-
mined which profile(s) to use, you can move on to Section D.5.3.
D.5.3 Creating FUNI Services

Create any FUNI services that you want and apply the FUNI and service profiles. For example:
myswitch:services funi-> new
Usage:
[[-id] <text>] Service ID (default: “”)
[-iatmif] <AtmIf> Input Port
[-timeslotstr] <text> Timeslots
[[-funi] <integer>] FUNI Profile Index (default: 0)
[[-service] <integer>] Service Profile Index (default: 0)
[[-traps] (enabled|disabled)] Traps (default: enabled)
[[-stats] (enabled|disabled)] Stats (default: disabled)
FramePlus Modules
myswitch:services funi-> new -iatmif 4c1 -timeslotstr 1 -name service_a
Configuring
notice: The newly created service ID is 4C1:00.
myswitch:services funi-> new -iatmif 4c1 -timeslotstr 2 -name service_b

myswitch:services funi-> new -iatmif 4c1 -timeslotstr 3 -name service_c

myswitch:services funi-> show

SvcId SvcInd TimeSlot Admin Funi Serv Signal Traps Stats Name
4C1:00 1108 1 up 0 0 nonexist enabled disabled service_a
4C1:01 1110 2 up 0 0 nonexist enabled disabled service_b
4C1:02 1112 3 up 0 0 nonexist enabled disabled service_c

As you can see in the Stats field, the collection of statistics is disabled, by default, at the ser-
vice level. It is enabled, by default, at the module level. If you want to collect statistics, you
must enable that functionality here at the service level as well:
myswitch:services funi-> modify -id 4C1:00 -stats enabled

notice: AAL5 Interval Statistics for Fabric IfIndex: 1108 have been reset


myswitch:services funi-> show

SvcId SvcInd TimeSlot Admin Funi Serv Signal Traps Stats Name
4C1:00 1108 1 up 0 0 nonexist enabled enabled service_a
4C1:01 1110 2 up 0 0 nonexist enabled enabled service_b
4C1:02 1112 3 up 0 0 nonexist enabled enabled service_c
D.5.3.1 Creating an Unstructured E1 Service

If you want to create an unstructured E1 service, perform the following steps:
1. Turn off CRC4 for the port using the following command:
interfaces dsx1 modify -hwportname <BNP> -linetype e1
2. Create a service profile with an access rate of 2048 Kbps as follows:
services profile service new -index <integer> -accrate 2048
3. Create a new service using the port from step 1 as the -iatmif, specifying the
timeslots (-timeslotstr) as either all or 0-31, and using the service profile
from step 2 for -service, as follows:
services funi new -iatmif <AtmIf> -timeslotstr <text> -service

<integer>

D.5.4 Creating FUNI PVCs

Now that the profiles and services have been created, you can create your FUNI PVCs and
apply the EPD/PPD profiles to the PVCs. For example:
myswitch:services funi pvc-> new

Usage:
[-id] <text> Service ID
[-fvpi] <integer> FUNI VPI
[-fvci] <integer> FUNI VCI
[[-oatmif] <AtmIf>] Output AtmIf
[[-ovpi] <integer>] Output VPI
[[-ovci] <integer>] Output VCI
[[-faupc] <UPC Index>] FAUPC (default: 0)
[[-afupc] <UPC Index>] AFUPC (default: 0)
[[-epdppd] <integer>] EPDPPD (default: 0)
To create a regular FUNI connection (one end on

NOTE the FramePlus network module and on end on
another network module), the -oport, -ovpi,
and -ovci must be specified. If this set of
parameters is not specified, a dangling PVC is
created. Port-to-port FUNI connections are not
supported.
FramePlus Modules
The default minimum VCI is 32 and the default
Configuring
NOTE maximum VCI is 63. If you are trying to create a
PVC on a FUNI service that uses the default
FUNI service profile, you are limited to this VCI
range when specifying the -fvci parameter.
myswitch:services funi pvc-> new -id 4C1:00 -fvpi 0 -fvci 40 -oatmif 4d1 -ovpi 0
-ovci 40 -name customer_a
-ovci 41 -name customer_b
-ovci 42 -name customer_c

myswitch:services funi pvc-> show

Input Output
SVCId FVPI FVCI AtmIf VPI VCI AtmIf VPI VCI Admin Name
4C1:00 0 40 4C1 0 40 4D1 0 40 up customer_a
4C1:00 0 40 4D1 0 40 4C1 0 40 up customer_a
4C1:01 0 41 4C1 16 41 4D1 0 41 up customer_b
4C1:01 0 41 4D1 0 41 4C1 16 41 up customer_b
4C1:02 0 42 4C1 32 42 4D1 0 42 up customer_c
4C1:02 0 42 4D1 0 42 4C1 32 42 up customer_c
D.5.5 Configuring FUNI SPVCs

You can also configure SPVCs (Smart Permanent Virtual Circuits) on a FramePlus network
module. An SPVC is a connection that spans multiple switch fabrics and looks like a PVC at
the local and remote endpoints with an SVC in the middle. If a link carrying an SPVC goes
down and there is an alternate route, then the end switch fabrics of the SPVC automatically
reroute the SPVC around the failed link.
However, the procedure for creating both SPANS SPVCs and PNNI SPVCs on a FramePlus net-
work module is slightly different than it is for other network modules. Section D.5.5.1
describes how to create a FUNI SPANS SPVC. Section D.5.5.2 describes how to create a FUNI
PNNI SPVC.
D.5.5.1 Creating a FUNI SPANS SPVC

To create a FUNI SPANS SPVC, perform the following steps:
ple, the switches are shark and tuna.
follows:
shark:services funi-> new -id 1a1:00 -iatmif 1a1 -timeslotstr 1


shark:services funi-> pvc

shark:services funi pvc-> new -id 1a1:00 -fvpi 0 -fvci 40
follows:
tuna:services funi-> new -id 1a1:00 -iatmif 1a1 -timeslotstr 1

tuna:services funi-> pvc

tuna:services funi pvc-> new -id 1a1:00 -fvpi 0 -fvci 40

FramePlus Modules
Configuring
SPANS Address from the source switch (shark) as the -srcswitchaddr value as
follows:

-srcswitchaddr 00000038.f21a333b -outatmif 1a1 -outvpi 0 -outvci 40

field as follows:

Local Remote State
1 1A1 0 40 0 1 up

Local Remote State
1 1A1 0 40 0 1 up


-srcswitchaddr 00000038.f21a3656 -outatmif 1a1 -outvpi 0 -outvci 40

Local Remote State
1 1A1 0 40 0 1 up

Local Remote State
1 1A1 0 40 0 1 up

D.5.5.2 Creating a FUNI PNNI SPVC

To create a FUNI PNNI PP SPVC, perform the following steps:
follows:
shark:services funi-> new -id 1a1:00 -iatmif 1a1 -timeslotstr 1

shark:services funi-> pvc

shark:services funi pvc-> new -id 1a1:00 -fvpi 0 -fvci 40
follows:
tuna:services funi-> new -id 1a1:00 -iatmif 1a1 -timeslotstr 1

FramePlus Modules
Configuring
tuna:services funi-> pvc
tuna:services funi pvc-> new -id 1a1:00 -fvpi 0 -fvci 40
6. Find the destination NSAP of the destination ATM interface (1A1 in this example)
on the destination switch as follows:


follows:

field as follows:

-----------------------------------------------------------------------------
1 1A1 0 40 0 require 5 up
0 0 40 0
complete.

D.6 Upgrading the FramePlus Network Module Software

Because the FramePlus network module has an on-board i960 processor, you can upgrade the
network module application software.
If you upgrade the switch software to

NOTE ForeThought 6.1, all of the FramePlus network
modules must also be upgraded to a compatible
FramePlus software application version. Please
see the following FORE web site for a hardware
and software compatibility matrix for FORE
network modules: http://www.fore.com/
products/swtch/dsatmnetmods.html
You should upgrade the switch SCPs before

NOTE upgrading the FramePlus network modules.
The method for upgrading the network module software is similar to that for upgrading the
switch software.
myswitch:hardware netmod application-> upgrade

Usage:
[-netmod] <Netmod> Netmod to upgrade
[-url] <URL> URL of form [(ftp|tftp)://]<host>/<full filepath>
FramePlus Modules
Configuring
[-netmod] <Netmod> The FramePlus network module on which you want to upgrade the software.
[-url] <URL> The upgrade information:

• ftp or tftp - the protocol used for the upgrade
• host - the IP address of the remote host on which the upgrade file resides
• full filepath - the full path name of the upgrade file

If you are using TFTP as the transfer protocol

NOTE (this is the default), the remote host on which the
upgrade file resides must be a tftpboot server. If
you are unsure of how to configure the bootp
server and the tftpboot server properly, see
Chapter 4 of the ATM Switch Installation and
Maintenance Manual.
For example:
myswitch:hardware netmod application-> upgrade 1a tftp://169.144.173.31/

N_FRAM_FR_ForeThought_00.01.00_1.5433.tar
Bringing netmod to BootConsole Mode
Downloaded 79 of 80 bytes
File download to netmod 1A successful

Now, Unfragmenting the netmod flash

Netmod upgrade successful
The network module is automatically reset if it was in the admin up state before the upgrade.
Otherwise, it remains in the admin down state and must be reset manually using hardware
netmod modify -admin up.
To display the current revision numbers of the application software, enter the following:
myswitch:hardware netmod application-> show -appversion
Netmod AppVersion
1A N_FR_ForeThought_1.4.00 FCS (1.46060)
To display the current revision number of the boot software, enter the following:
myswitch:hardware netmod application-> show -bootversion

Netmod BootVersion
4B 1.0.0

APPENDIX E RMON
This appendix provides a basic description about Management Information Base (MIB)
objects and the implementation of RMON with extensions for an ATM network on Enterprise
switches. RMON stands for “remote monitor” and is a specification for the Simple Network
Management Protocol (SNMP) management interface for remote network monitors imple-
mented either as a stand-alone probe or as an agent residing on a switch.
This appendix contains the following information:
• Section E.1 - MIB Overview - Provides a brief overview of network management
systems and the relationship between SNMP and Management Information Base
(MIB) objects in the TCP/IP network.
• Section E.2 - SNMP Overview - Describes the common operations performed by
SNMP to support the exchange of management information between networked
devices.
• Section E.3 - RMON Overview - Describes the RMON specification and provides
a snapshot of the associated table definitions supported in this release.
RMON
ATM Switch Network Configuration Manual E-1

RMON
E.1 MIB Overview

To promote interoperability, systems must adhere to a common framework and a common
language, called a protocol. For network management, that protocol is SNMP. The most com-
mon managing system is called a network management system (NMS). Managed devices can
include hosts, servers, or network components such as switches.
An example of an NMS would be HP OpenView, which provides a management application
called Network Node Manager (NNM) for use in managing TCP/IP networks and network
devices that support SNMP. This application runs under the HP OpenView Windows (OVW)
graphical user interface. HP OpenView provides configuration, performance, and fault man-
agement support for multi-vendor networks.
In a managed device, specialized software modules called agents access information about the
device and make it available to the NMS. Managed devices maintain values for a number of
variables and report those, as required, to the NMS. For example, an agent might report such
data as the number of bytes and packets in and out of the device. In the TCP/IP network man-
agement framework, each variable is referred to as a managed object, which is anything that
an agent can access and report back to the NMS.
All managed objects are contained in the MIB, which is a database of the managed objects. The
managed objects, or variables, can be set or read to provide information on network devices
and interfaces. An NMS can control a managed device by sending a message to an agent of
that managed device requiring the device to change the value of one or more of its variables.
E-2 ATM Switch Network Configuration Manual

RMON
E.2 SNMP Overview

MIB variables can be accessed through SNMP, an application-layer protocol designed to facili-
tate the exchange of management information between networked devices. The SNMP system
consists of three parts: the SNMP manager, the SNMP agent, and MIBs.
• The SNMP manager can be part of a network management system (NMS), and
the SNMP agent can reside on a networking device such as a switch. You can
compile MIBs with the network management software. If SNMP is configured on
a target device such as a FORE ASX switch, the SNMP agent responds to the MIB-
related queries being sent by the NMS.
• The SNMP agent gathers data from the MIB. Then, the agent can send traps, or
notifications of certain events, to the manager. The SNMP manager uses informa-
tion in the MIB to perform the operations described in Table E.1.
Table E.1 - Common Operations Supported by SNMP
Operation Description
get A management station retrieves a value from a specific variable.
set A management station updates a value in a specific variable in an agent.
trap An unsolicited message sent by an SNMP agent to an SNMP manager indicat-

ing that some event has occurred.
RMON

RMON
E.3 RMON Overview

The original RMON specification known as RMON-I specifies a collection of SNMP groups
designed primarily for monitoring ethernet networks. This specification has been augmented
by the addition of MIB objects for monitoring Token Ring networks, MIB objects for monitor-
ing the protocol traffic above the MAC level (level 3-7 in the OSI model), MIB objects for mon-
itoring ethernet switches, and objects for monitoring ATM networks.
This implementation of RMON invokes portions of the following standards documents:
• Remote Network Monitoring MIB, RFC 1757, for simple threshold monitoring
and event logging and notification
• Remote Monitoring MIB Extensions for ATM Networks, AF-NM-TEST-0080.000,
for basic ATM statistics
The RMON probe monitors address-based network traffic.
E.3.1 RMON Table Definitions

The following tables from RMON are implemented in ForeThought 6.1 and greater:
• portSelect - port selection; this group implements the switch port aggregation
for the RMON MIB. This group replaces the columnar data-source and arbitrary
integer index used in all the other RMON MIBs. This group consists of a “port
select group(ing) table” called the portSelGrpTable and a “port selection”
table called the portSelTable.
• atmStats - basic statistics; accumulates individual host traffic across a port select
instance.
The following tables from RMON are not implemented because of their resource intensive
nature:
• atmHost - ATM statistics tabulated and sorted by host address.
• atmMatrix - ATM statistics tabulated and sorted by host-to-host conversations
and additionally sorted by TopN (top number) of host-to-host conversations.
The following tables from RMON-I RFC 1757 are implemented:
• Alarm - allows simple threshold monitoring of switch MIB variables. The alarm
table uses the event table for generating and transmitting notifications when a
threshold is crossed.
• Event - the event table controls the generation and notification of events from
this device. This group consists of the eventTable and the logTable.

RMON
E.3.1.1 The portSelect Group

The port select group implements the data source used by the groups in the RMON MIB. It is
designed to enable aggregation of one or more physical or logical ATM ports into one data
source from which statistics can be gathered.
The ATM Port Selection MIB:
fore
experimental(xx)
atmRmonMIBObjects(1)
portSelect(1)
portSelGrpTable(1)
portSelGrpEntry(1) Index {portSelGrpIndex}
portSelGrpIndex(1)
portSelGrpIndexDescr(2)
portSelGrpCreateTime(3)
portSelGrpOwner(4)
portSelGrpStatus(5)
portSelTable(2)
portSelEntry(1) Index {ifIndex}
portSelCollectGroup(1)
portSelCreateTime(2)
portSelOwner(3)
portSelStatus(4)
RMON

RMON
E.3.1.2 The atmStats Group

The ATM statistics group counts the total amount of host addressed traffic on behalf of one or
more portSelectGroups. The statistics group consists of two tables: a statistics control table
that can be created and configured by the network management station and a statistics data
table that will only exist when the associated control table row status if active. The control
table is indexed by the portSelGrpIndex from the port select group.
The ATM Statistics Group MIB:
fore
experimental(xx)
atmRmonMIBObject(1)
atmStats(2)
atmStatsControlTable(1)Index{portSelGrpIndex}
atmStatsControlEntry(1)
atmStatsControlDropEvents(1)
atmStatsControlOwner(2)
atmStatsControlStatus(3)
atmStatsTable(2)
atmStatsEntry(1)Index{portSelGrpIndex, atmStatsSClass}
atmStatsSClass(1)
atmStatsCreateTime(2)
atmStatsCells(3)
atmStatsCellsRollovers(4)
atmStatsHCCells(5)
atmStatsNumCallAttempts(6)
atmStatsNumCalls(7)
atmStatsConnTime(8)

RMON
E.3.1.3 The RMON Alarm Group

The RMON Alarm group is a component of RMON-I and is defined in RFC 1757. The Alarm
group periodically takes statistical samples from variables defined in the probe and compares
them to thresholds that have been configured. If a sample crosses the threshold boundary, an
alarm event is generated and forwarded to the event group.
The RMON Alarm Group MIB:
mib-2
rmon (16)
alarm(3)
alarmTable(1)
alarmEntry(1) Index {alarmIndex}
alarmIndex(1)
alarmInterval(2)
alarmVariable(3)
alarmSampleType(4)
alarmValue(5)
alarmStartupAlarm(6)
alarmRisingThreshold(7)
alarmFallingThreshold(8)
alarmRisingEventIndex(9)
alarmFallingEventIndex(10)
alarmOwner(11)
alarmStatus(12)
RMON

RMON
E.3.1.4 The Event Group

The Event group controls the generation of events from MIB groups in the switch. The
RMON-I alarm group is the primary user of this table. The Event group enables the customiz-
ing of events triggered via this table. The Event group is composed of two tables, an event
control table called the eventTable and a event data table the called logTable. The event
control table contains rows of user configured events. MIB instances sending events to the
event group use the control table index or key value to identify the row to which an event is
sent. The event data table contains a log of events sent to the control table.
The Event Group MIB:
mib-2
rmon(16)
event(9)
eventTable(1)
eventEntry(1) Index {eventIndex}
eventIndex(1)
eventDescription(2)
eventType(3)
eventCommunity(4)
eventLastTimeOut(5)
eventOwner(6)
eventStatus(7)
logTable(2)
logEntry(1) Index {logEventIndex, logIndex}
logEventIndex(1)
logIndex(2)
logTime(3)
logDescription(4)

Configuring IMA
Modules
Universal IMA network modules are multi-service network modules containing eight DS1 or
E1 ports that support ATM UNI and Inverse Multiplexing for ATM (IMA). IMA provides a
cost-effective way to increase WAN ATM bandwidth in increments of T1/E1. The universal
IMA network modules also provide all of the capabilities of the Series D network modules.
This appendix provides an overview of a universal IMA network module and provides con-
figuration examples. Topics that are covered include the following:
• Section F.1 - IMA Overview
• Section F.2 - Configuring IMA
• Section F.3 - Testing an IMA Group
• Section F.4 - Upgrading the IMA Network Module Software
ATM Switch Network Configuration Manual F-1

Configuring IMA Modules
F.1 IMA Overview

IMA is a method for aggregating multiple physical media into a single, logical link so that it
appears as though all ATM-related functionality (e.g., signalling, creation of VPs and VCs) is
taking place over a single, physical medium. IMA is useful for transmitting traffic from one
source over multiple low-speed physical links when the source bandwidth requirements are
greater than any one low-speed physical link. This method of aggregation is illustrated in
Figure F.1.
Link a 4, 1 Re-aggregated
Tx Cell Sequence Cell Sequence
Link b 5, 2
6, 5, 4, 3, 2, 1 6, 5, 4, 3, 2, 1
Link c 6, 3
universal IMA network module
In the transmit direction, cells In the receive direction, cells
are distributed across the links are reconstructed into a
in round-robin sequence. single ATM stream.
Figure F.1 - Multiple Physical Links Aggregated into One Logical Link
F.1.1 Links and Groups

On a universal IMA network module, a link is the same as a port. These links can be assigned
together logically as a group. A link may only be assigned to one group at most. It cannot be
assigned to multiple groups simultaneously. Links may be assigned to a group in any order
and need not be contiguous. For example, a group could contain ports 1, 3, and 8; or ports 2
and 5. However, the physical links that can participate in an IMA group must be links that
reside on the same network module.
A group may contain from 0 to 8 physical DS1 or E1 links. As a logical ATM interface, a group
may be used as an endpoint for ATM connections. The bandwidth of a group is approximately
the sum of bandwidth of the links that belong to that group. As many as eight IMA groups,
with each group containing one link, may reside on the same network module. However, this
is not a typical configuration.
F-2 ATM Switch Network Configuration Manual

F.1.2 Benefits
Configuring IMA
The universal IMA network module provides the following important benefits.
Modules
F.1.2.1 Scalability
The universal IMA network module allows you to increase your network capacity by incre-
mentally adding to your existing, leased lines. For example, if you already have a leased DS1
line, but require a small amount of additional bandwidth, you can lease a second DS1 line and
use the universal IMA network module to approximately double the bandwidth of your con-
nection. The second line can be added without disrupting the existing traffic. This provides a
cost-effective alternative to upgrading to a DS3 line.
F.1.2.2 Load-balancing
The universal IMA network module is load-balanced so that when several physical DS1 or E1
lines are aggregated together in an IMA group, the services running on the link are main-
tained even if one or more of the individual links in the group goes down. Additionally, you
can add physical links to or delete physical links from an existing group without disruption to
the services that are already running on that group.
F.1.2.3 ATM Connection Aware IMA

If one of the links in an IMA group goes down, the software checks the resulting bandwidth in
the IMA group. If the new operational bandwidth of the IMA group (after accounting for the
loss of bandwidth due to the link going down) is greater than or equal to the sum total band-
width being used by all the connections on that IMA group, then no action is taken.
However, if the IMA group has become overbooked (i.e., does not have enough bandwidth to
maintain the current connections), then the software starts releasing lower priority connec-
tions to maintain the QOS requirements of the higher priority connections. Currently, the pri-
ority scheme is as follows: ABR SVCs are released first, followed by nrtVBR SVCs, and, finally,
the rtVBR SVCs and CBR SVCs. UBR SVCs are ignored and left intact because they are not
allocated bandwidth anyway. The software releases only as many connections as required to
bring the IMA group out of the overbooked state. If the IMA group is still in an overbooked
state and all of the SVCs have been released, then the software starts releasing the SVPs. ABR
SVPs are released first, followed by nrtVBR, rtVBR, and CBR SVPs (in that order). Again, UBR
SVPs are ignored. None of the permanent connections (PVCs and PVPs) are ever dropped.
F.1.2.4 Full Series D Capabilities

The universal IMA network module provides the same functionality as a Series D DS1 or E1
network module. All ATM layer services are available, such as PNNI, SVCs, SPVCs, LANE,
MPOA, etc. Each port can function as either an ATM UNI or as part of an IMA group.

F.2 Configuring IMA

A universal IMA network module functions as a Series D DS1 or E1 network module. To take
advantage of the IMA features, you must configure IMA on a universal IMA network module
using the following steps in this order:
1. Create a group and assign one or more links to the group.
2. Create any PVPs, PVCs, or SPVXs on the interface as you normally would.
These steps are described in detail in the following sections.
F.2.1 Creating a Group and Assigning Links

First, you need to create a group using the interfaces ima group new ? command.
To display the command usage, you must enter a

NOTE ? as part of the command. If you enter the
interfaces ima group new command
without specifying a ?, the network module
immediately creates a new group with no links.
For example, to create a group that will contain three links, enter the following, using the
-imalinks option to assign the links to the group:
myswitch:interfaces ima group-> new ?

[[-index] <ImaIf>] Index (default: IMA3)
[[-symmetry] <symmetry>] Symmetry (default: symOper)
[[-minnumtxlinks] <integer>] Min Num Tx Links (default: 1)
[[-minnumrxlinks] <integer>] Min Num Rx Links (default: 1)
[[-netxclkmode] <clockmode>] NE Tx Clock Mode (default: ctc)
[[-tximaid] <integer>] Tx Ima Id (default: 3)
[[-txframelength] <framelength>] Tx Frame Length (default: m128)
[[-diffdelaymax] <integer>] Diff Delay Max (default: 25)
[[-alphavalue] (1|2)] Alpha Value (default: 2)
[[-betavalue] (1|2|3|4|5)] Beta Value (default: 2)
[[-gammavalue] (1|2|3|4|5)] Gamma Value (default: 1)
[[-imalinks] <text>] List of IMA Links (default: “”)
myswitch:interfaces ima group-> new -index IMA1 -imalinks 1A1:1A2:1A3

You cannot modify a group. Once it has been

NOTE created, you must delete it and recreate it if you
Configuring IMA
wish to change anything.
Modules
Links may be added either at the time you create
the group by using the -imalinks option, or by
using the interfaces ima link new
command as shown in Section F.2.1.1. If using
the -imalinks option, the links must be
delimited with a colon; e.g., -imalinks
1a1:1a2:1a3.
F.2.1.1 Assigning Additional Links to the Group

If, at a later time, you decide to assign some additional links to the group that you previously
created called IMA1, you can easily do this as follows:
myswitch:interfaces ima link-> new

Usage:
[-index] <text> Index
[-groupindex] <ImaIf> Group Index
myswitch:interfaces ima link-> new -index 1A4 -groupindex IMA1

myswitch:interfaces ima link-> new -index 1A5 -groupindex IMA1
Alternatively, you could simply enter the following:
myswitch:interfaces ima link-> new 1A4 IMA1

myswitch:interfaces ima link-> new 1A5 IMA1

F.2.1.2 Deleting Links from the Group

If, at a later time, you decide to remove some links from the group that you previously cre-
ated, you can easily do this.
If you delete a link from a group, and the deletion does not cause the IMA group to go into an
overbooked state, then there is no loss of service. However, if the link being deleted is the cur-
rent source of the switch distributed timing, then the software selects a new source and recon-
figures all of the ports on the switch (including the ports belonging to the IMA group) with the
new timing source. This timing reconfiguration can cause a temporary loss of service. Once
the ports are configured, there is no further loss of service.
Note that this is different from deleting the Timing Reference Link (TRL) of the IMA group.
The user should be able to delete the TRL, just like any other link in the IMA group, and there
will be no loss of service.
If you attempt to delete a link from an IMA group and the resultant configured bandwidth
would fall below the total bandwidth allocated to the permanent connections (PVCs and
PVPs), your request will be rejected. An error message to this effect will be output on the con-
sole. (The configured bandwidth of an IMA group is approximately equal to the number of
links in an IMA group multiplied by the cell rate of each link.)
To delete a link from an IMA group, enter the following parameters:
myswitch:interfaces ima link-> delete

Usage:
[-index] <Dsx1If> Index
[[-groupindex] <ImaIf>] Group Index
[[-reldelay] <integer>] Rel Delay
[[-netxstate] <linkstate>] NE Tx State
[[-nerxstate] <linkstate>] NE Rx State
[[-fetxstate] <linkstate>] FE Tx State
[[-ferxstate] <linkstate>] FE Rx State
[[-nerxfailurestatus] <linkfailurestatus>] NE Rx Failure Status
[[-ferxfailurestatus] <linkfailurestatus>] FE Rx Failure Status
[[-txlid] <integer>] Tx Link ID
[[-rxlid] <integer>] Rx Link ID
[[-testpattern] <integer>] Test Pattern
[[-testprocstatus] <procstatus>] Test Proc Status
[[-imaviolations] <integer>] IMA Violations
[[-oifanomalies] <integer>] OIF Anomalies
[[-neseverroredsec] <integer>] NE Sev Errored Sec
[[-feseverroredsec] <integer>] FE Sev Errored Sec
[[-neunavailsec] <integer>] NE Unavail Sec
[[-feunavailsec] <integer>] FE Unavail Sec

[[-netxunusablesec] <integer>] NE Tx Unusable Sec

[[-nerxunusablesec] <integer>] NE Rx Unusable Sec
Configuring IMA
[[-fetxunusablesec] <integer>] FE Tx Unusable Sec
Modules
[[-ferxunusablesec] <integer>] FE Rx Unusable Sec
[[-netxnumfailures] <integer>] NE Tx Num Failures
[[-nerxnumfailures] <integer>] NE Rx Num Failures
[[-fetxnumfailures] <integer>] FE Tx Num Failures
[[-ferxnumfailures] <integer>] FE Rx Num Failures
[[-txstuffs] <integer>] Tx Stuffs
[[-rxstuffs] <integer>] Rx Stuffs
myswitch:interfaces ima link-> delete -index 1A2 -groupindex IMA1

myswitch:interfaces ima link-> delete -index 1A3 -groupindex IMA1
Alternatively, you could simply enter the following:
myswitch:interfaces ima link-> delete 1A2 IMA1

myswitch:interfaces ima link-> delete 1A3 IMA1

F.2.2 Creating PVPs, PVCs, or SPVXs

At this point, group IMA1 contains links 1A1, 1A2, 1A3, 1A4, and 1A5. Group IMA1 is like the
remaining ports 1A6 through 1A8 on the network module, except it now has five times the
bandwidth of each of those ports. You can now create connections on IMA1 as you would on
any other port.
You can create connections from group to group

NOTE or from group to port on a single universal IMA
network module. You can also create them from
group to group or from group to port between a
universal IMA network module and any other
IMA or non-IMA network module.
Creating connections on an IMA group is done

NOTE exactly the same way as it is done on the other
ATM interfaces (AtmIfs). The only difference is
that for regular ATM ports, the ATM interface
name is in the Board-Netmod-Port (BNP) format
(e.g., 1A1). For IMA groups, the ATM interface
name is the name of IMA group (e.g., ima1, ima2,
etc).
The following is an example of creating a PVC on an IMA group:
myswitch:connections channel-> new -iatmif ima1 -ivpi 0 -ivci 100 -oatmif 1d1 -ovpi 0
-ovci 100 –name Ima1_1d1
myswitch:connections channel-> show -iatmif ima1

Input Output
AtmIf VPI VCI AtmIf VPI VCI UPC Protocol Name
IMA1 0 5 1CTL 0 37 0 fsig N/A
IMA1 0 14 1CTL 0 36 0 spans N/A
IMA1 0 15 1CTL 0 35 0 spans N/A
IMA1 0 16 1CTL 0 38 0 fsig N/A
IMA1 0 18 1CTL 0 111 0 rcc N/A
IMA1 0 100 1D1 0 100 0 pvc ima1_1d1

The number of outgoing rate-controlled (shaped

NOTE or guaranteed) connections that you can create
Configuring IMA
on an IMA group is limited by the
-mintxlinks parameter of the IMA group. For
Modules
example, if the -mintxlinks parameter of an
IMA group on a DS1 IMA netmod is set to 1,
then you can create a maximum of one DS1
worth of rate-controlled (shaped or guaranteed)
connections. This provides the rate-controlled
connections with the highest priority.
The following is an example of creating a PVP on an IMA group:
myswitch:connections path through-> new -iatmif ima1 -ivpi 5 -oatmif ima1 -ovpi 5
The following is an example of creating a point-to-point PNNI SPVP on an IMA group:
myswitch:connections smartpath-> new -index 1 -callingatmif ima1 -callingvpi 1

-calledatmaddr 47000580ffe1000000f21c378a0020481c378a00
The following is an example of creating a point-to-point PNNI SPVC on an IMA group:
myswitch:connections smartchannel pnni pp-> new -index 1 -callingatmif ima1 -callingvpi

0 -callingvci 100 -calledatmaddr 47000580ffe1000000f21c378a0020481c378a00

F.3 Testing an IMA Group

You can test the connectivity of a link within an IMA group. A test pattern, which is specified
by the -testpattern option, is sent over one link, which is specified by the -testlink
option. This pattern is expected to be looped over all of the other links in the group at the far
end. If no specific test link and pattern are specified by the user, the switch chooses a default
pattern and chooses the first link.
You can modify these test parameters as follows:
myswitch:interfaces ima group test-> modify ?

[-testindex] <ImaIf> Index
[[-testlink] <Dsx1TestLink>] Test Link
[[-testpattern] <integer>] Test Pattern
[[-testprocstatus] <procstatus>] Test Proc Status
F.3.1 Choosing a Test Link

The test link must be a DSX1 (DS1 or E1) port on a universal IMA network module and must
be one of the links in the IMA group being modified. The -testlink option can be set to any
to indicate that the switch should choose the test link when a test is initiated. This is the
default setting.
F.3.2 Choosing a Test Pattern

The default -testpattern is –1, which indicates that the switch should choose a test pattern
when a test is initiated on the group. The valid values for -testpattern are from –1 to 255,
inclusive.
F.3.3 Starting or Stopping a Test Procedure

The -testprocstatus option is used to start or stop a test. Setting -testprocstatus to
operating starts the test and setting it to disabled stops the test. The default is disabled. If
any link fails, the test continues. At the end of the test, a value of linkfail is displayed if one
or more links failed during the course of the test.
F - 10 ATM Switch Network Configuration Manual

F.3.4 Test Example
Configuring IMA
The following is an example of how to start a test:
Modules
msywitch:interfaces ima group test-> modify ima2 -testlink 1a4 -testpattern 180
-testprocstatus operating
myswitch:interfaces ima group test-> show ima2

Test Test Test
Group Link Pattern Proc Status
IMA2 1A4 180 operating
The following is an example of how to stop a test:
msywitch:interfaces ima group test-> modify ima2 -testprocstatus disabled
myswitch:interfaces ima group test-> show

Test Test Test
Group Link Pattern Proc Status
IMA2 1A4 180 disabled
ATM Switch Network Configuration Manual F - 11

F.4 Upgrading the IMA Network Module Software

Because the universal IMA network module has an on-board ATMizer processor, you can
upgrade the network module application software.
When using universal IMA network modules in

NOTE a switch, ForeThought 6.1.x or greater is required
on the switch SCP(s). Also, all of the IMA
network modules must be running a compatible
IMA software application version. If the IMA
network modules and the switch SCP(s) are not
running corresponding releases, the protocols
are incompatible and an error message results.
Please see the following FORE web sites for
hardware and software compatibility matrices
for FORE network modules:
http://www.fore.com/products/netmods/
netmods_hw_pcm.html
or
http://www.fore.com/products/netmods/
netmods_sw_pcm.html
You should upgrade the switch SCPs before

NOTE upgrading the universal IMA network modules.
The method for upgrading the network module software is similar to that for upgrading the
switch software.
myswitch:hardware netmod application-> upgrade

Usage:
[-netmod] <Netmod> Netmod to upgrade
[-url] <URL> URL of form [(ftp|tftp)://]<host>/<full filepath>
[-netmod] <Netmod> The universal IMA network module on which you want to upgrade the software.
[-url] <URL> The upgrade information:

• ftp or tftp - the protocol used for the upgrade
• host - the IP address of the remote host on which the upgrade file resides
• full filepath - the full path name of the upgrade file

If you are using TFTP as the transfer protocol

NOTE (this is the default), the remote host on which the
Configuring IMA
upgrade file resides must be a tftpboot server. If
you are unsure of how to configure the bootp
Modules
server and the tftpboot server properly, see
Chapter 4 of the ATM Switch Installation and
Maintenance Manual.
If you manually admin the network module down prior to the upgrade (using hardware
netmod modify -admin down), you must manually admin it up after the upgrade (using
hardware netmod modify -admin up). However, if you upgrade the network module
while it is in the admin up state, the software automatically admins it down, performs the
upgrade, and then admins it back up without any additional user intervention.
myswitch:hardware netmod-> upgrade 1a tftp://169.144.173.31/

N_IMA_ForeThought_00.01.00_1.5433.tar
Bringing netmod to BootConsole Mode


Netmod upgrade successful
To display the current revision number of the application software, enter the following:
myswitch:hardware netmod application-> show -appversion

Name AppVersion
1A N_IMA_ForeThought_00.01.00 (1.5433)
To display the current revision number of the boot software, enter the following:
myswitch:hardware netmod application-> show -bootversion

Name BootVersion
1A 1.0.0
ATM Switch Network Configuration Manual F - 13


ForeThought 6.1.x enhances ASX-4000 redundancy with the following features:

• line/port level SONET 1+1 unidirectional non-revertive automatic protection
switching (APS)
ASX-4000 Redundancy
• passive fabric protection
• passive port card protection
The ASX-4000 can only be configured for redundant

NOTE operations with Revision E or higher fabrics. Fabric
revisions prior to E do not support the redundancy
feature.
This appendix provides an overview of SONET APS and the ASX-4000 redundancy features
that are implemented. Topics that are covered include the following:
• Section G.1 - Overview of SONET APS
• Section G.2 - Terminology
• Section G.3 - Redundant Switch Configurations
• Section G.4 - Redundant Link Naming Convention
• Section G.5 - Configuring Working and Protection Interfaces
• Section G.6 - Partner Synchronization
• Section G.7 - Loopbacks on Redundant Ports
ATM Switch Network Configuration Manual G-1

ASX-4000 Redundancy
G.1 Overview of SONET APS

ForeThought 6.1 supports SONET linear 1+1 automatic protection switching (APS) on an
ASX-4000. 1+1 SONET APS is provided when Revision E fabrics are configured for APS and
two port cards of the same type are installed and configured for redundancy. 1+1 SONET APS
allows for a working link to backed up by a protection link, with the working link data being
replicated onto the protection link. In the event of a failure on the working link, the far end
receiver switches over to receiving data from the protection link within 50 milliseconds.
ForeThought 6.1.x provides a passive 1+1 fabric and port card protection scheme between
working and corresponding protection fabrics/port cards through the implementation of
user-initiated failover routines (fabric/port card hot-swaps).
G.1.1 Line/Port Protection

In an effort to minimize the time a network is out-of-service due to line /port failures,
Telcordia has defined schemes for Automatic Protection Switching (APS) providing high
speed detection of failure conditions on SONET facilities and the automatic failover of failed
working links to standby protection facilities.
Working Links
TX RX
RX TX
TX RX
RX TX
Protection Links
Figure G.1 - Linear APS Failover to the Protection Link upon Detection of SONET Line Failure
G.1.1.1 Linear APS

The ASX-4000 supports the SONET Linear 1+1 APS architecture at the line/port level. In the
1+1 architecture, the transmitted signal is continuously bridged to both working and protec-
tion equipment. The receiver monitors both transmitted signals and chooses the signal that
gives better quality as the one from which it selects traffic. The other signal is discarded at the
receiver. The 1+1 architecture does not allow for extra traffic to ride on the protection port. The
ASX-4000 conforms to Telcordia GR-253-CORE.
G-2 ATM Switch Network Configuration Manual

ASX-4000 Redundancy
G.1.1.2 Unidirectional Mode

The ASX-4000 supports unidirectional switching. In the unidirectional mode, each receiver
autonomously choose whether it will select traffic from the working or the protection channel.
Although a switch between these channels will be signalled by the receiver to the transmitter,
the transmitting end is not required to read this signalling and therefore does not necessarily
know which of its two transmitted signals is actually being used by the receiver. It is not
required to switch its own receiver when the other end switches. Thus it is possible on a link
where transmission occurs in both directions for the receiver at each end of the link to have
selected and be operating from a different physical transmission link.
ASX-4000 Redundancy
G.1.1.3 Failure Recovery / Switchover Performance
ASX-4000 ports configured for SONET Linear 1+1 APS can be configured for non-revertive
manual recovery mechanism or revertive automatic recovery mechanism. Non-revertive man-
ual recovery specifies that in the event of a failure on the working link and the subsequent
failover to the protection link, traffic is maintained on the protection link even after the work-
ing link has recovered from its original failure condition. Manual intervention is required for
recovery purposes, to direct traffic back to the working line/port, or port card. Implementing
non-revertive recovery avoids an unnecessary service disruption (to switch back to the work-
ing channel as in the revertive mode).
G.1.1.4 Manual and Automatic Protection

Manual protection is useful when traffic must be redirected in order to perform fiber or port
upgrades and maintenance. The ASX-4000 provides AMI, embedded WEB interface and
SNMP “maintenance commands” supporting a manual failover from working to protection
links on both a port-by-port and port card-wide basis.
Automatic line/port protection is provided to guard against random events that may be a
result of equipment or infrastructure failures that would otherwise result in loss or degrada-
tion of traffic in the network. The following condition initiates automatic protection:
• Hard Failure Conditions: Hard failure conditions include situations that result in
immediate traffic loss or signal failure in the network. Hard failure conditions
include the receipt of loss of signal (LOS), loss of frame (LOF), or alarm indication
signal (AIS-L).
G.1.1.5 Protection Failover Performance

Ports configured for SONET linear 1+1 APS complete failover in less than 50 ms upon the
detection and initiation of the failover action. The ASX-4000 supports failover regardless of
whether the original port card has been re-seated or replaced.

ASX-4000 Redundancy
G.1.1.6 Configuration and Management

The ASX-4000 supports the configuration of Linear 1+1 APS working and protection links
between end points using two fiber pairs (i.e., working pair and protection pair).
Working Port Car d Working Port Car d

Near-End Device Far-End Device
TX
RX Port 1
Port 1
RX
Fiber Pair X TX
Port N
Port N Fiber Pair Y
TX RX
Port 1 Port 1
RX TX
Port N
Port N
Prot ection Po rt Card Protection Po rt Card

Near -End Device Far-End Device
Figure G.2 - Port Redundancy
The ASX-4000 also supports SONET linear 1+1 APS on a port card-wide basis where all lines/
ports associated with both the working and protection physical port cards are placed into a
1+1 protected mode of operation.
1+1 Port Card Pair
Port 1 Tx 1+1 Port Pair Tx Port 1

Rx Rx
Port N Tx 1+1 Port Pair Tx Port N

Rx Rx
Working Port Card Protection Port Card

“Port Card 1AB” “Port Card 2AB”
Figure G.3 - Port Card Redundancy

ASX-4000 Redundancy
The ASX-4000 supports APS between Revision E fabrics and working and protection port
cards of the same series and type (i.e. Series 1: 8 Port 622 Mbps SMIR and Series 1: 8 Port 622
Mbps SMIR; or Series 2: 8 Port 622 Mbps SMIR and Series 2: 8 Port 622 Mbps SMIR). However,
port cards that are not the same series and type may exist between near-end and far-end
switches. The working port card and its partner protection port card is fixed and is dependent
on the slot that the working port card is installed. The interface group of the working port card
should be the same for the protection port card. For example, if 1AB is the working port card,
2AB is its protection port card. If 3CD is the working port card, 4CD would be its protection
port card. The protection fabric is also fixed and is dependent on the fabric that is being pro-
tected (working). See Section G.5 for more information.
ASX-4000 Redundancy
The ASX-4000 is compatible with 3rd party SONET and SDH equipment including Add/Drop
Multiplexers and ATM switches so long as they are compliant with the mandatory require-
ments specified under section 5.3, Linear 1+1 APS Architecture, in Telcordia GR-253-CORE.

ASX-4000 Redundancy
G.1.2 Fabric and Port Card Protection

Figure G.4 shows a protected line/port/port card/fabric configuration for the ASX-4000 (half
of the switch is shown). Fabrics 1 and 2 comprise of an active/standby protection pair with
fabric 1 configured as the working fabric, and fabric 2 as the protection fabric. Port cards 1AB
and 1CD are similarly designated as the working port cards, with port cards 2AB and 2CD,
their respective protection elements. All received traffic destined for an output on port cards
1AB or 1CD is forwarded to fabric 1 and fabric 2.
PORT CARD 1C/D PORT CARD 2A/B
PORT CARD 1A/B PORT CARD 2C/D
FABRIC 1 FABRIC 2 Fabrics 1 and 2 are a

Working/Protection pair
40GBPS BACKPLANE BUS
Primary Transmit Cell Bus
Secondary Transmit Cell Bus
Figure G.4 - 1+1 Data Path Protection (Half of Switch Is Shown)

ASX-4000 Redundancy
Each fabric transmits outbound cells onto both a primary and a secondary transmit cell bus,
with one bus connecting to a working port card and the other to its corresponding protection
card. At any time, a port card is actively listening to either its primary or secondary transmit
bus. If a working fabric or port card is hot-swapped, the corresponding protection element can
take over in its place. If a port failure or fiber cut occurs, the receiver can detect the failure and
switch to the standby circuit within 50 milliseconds. Such rapid protection switching provides
transparent recovery by maintaining the connection with minimal cell loss.
G.1.2.1 Manual and Active Protection
ASX-4000 Redundancy
The ASX-4000 supports manual port card/fabric redundancy. Manual protection includes
failovers initiated by AMI, WEB and SNMP commands as well as failovers that occur when
port cards/fabrics are removed or unseated from the ASX-4000.
The ASX-4000 is capable of detecting and completing port card/fabric failover in less than 50
ms. Port card switchover is non-revertive while fabric switchover can be configured for rever-
tive or non-revertive modes.
G.1.3 Timing Source

If a port is part of an APS redundant group and if it is configured as a timing source, a
switchover of this port will switch the timing source to its partner port. For example, if 1A1
and 2A1 are part of an APS group where 1A1 is active and configured as the primary clock,
upon switchover from 1A1 to 2A1, timing will be switched to 2A1 where 2A1 will become the
new timing source.
It is recommended to configure primary and

NOTE secondary switch clock sources only on working
ports.

ASX-4000 Redundancy
G.2 Terminology
Throughout the documentation, certain terms are used in reference to 1+1 redundancy. The
following terms are defined as follows:
Term Description
working Used in the 1+1 protection sense to refer to the logical state of
various switch components (i.e., lines/ports, fabrics, port cards)
that are designated as the principal path for carrying traffic
between the head-end and tail-end equipment.
protection Used in the 1+1 protection sense to refer to the logical state of
various switch components (i.e. lines/ports, fabrics, port cards)
that are designated as the redundant or failover path for carry-
ing traffic between the head-end and tail-end equipment.
active Refers to the link or component that is actually carrying data at
any given time.
standby Refers to the operational state of an ASX-4000 link or compo-
nent that is functioning as a hot backup in the event that the
active component should fail.
link An entity that defines a topological relationship including avail-
able transport capacity between two nodes in different subnet-
works or devices. The term link is specifically used to define the
physical transmit or receive data path of a specific port.
switchover Refers to the transition from protection to working components
as a result of a manually, forced, or automatically initiated
action.

ASX-4000 Redundancy
G.3 Redundant Switch Configurations

The ASX-4000 can be configured through software for different operating modes depending
on the degree of hardware redundancy best suited to the application. The following sections
describe the supported configurations.
G.3.1 20 Gbps Fully Redundant Configuration

For the most critical applications, the ASX-4000 can be configured as a fully redundant 20
ASX-4000 Redundancy
Gbps switch with full 1+1 data path redundancy. In this configuration, ports, port cards, and
fabrics are protected on a 1+1 basis (see Figure G.5). Redundant switches are part of a cost-
effective restoration strategy for voice and multi-service network providers with protected
SONET/SDH infrastructure.
5 Gbps Port Card 1AB

Fabric 1
40 Gbps
5 Gbps Port Card1CD
40 Gbps
Backplane
40 Gbps Fabric 2
5 Gbps Port Card 2CD
Working
5 Gbps Port Card 3AB Protection

40 Gbps Fabric 3

40 Gbps Fabric 4
Figure G.5 - ASX-4000 20 Gbps Fully Redundant Configuration

ASX-4000 Redundancy
G.3.2 30 Gbps Mixed Configuration

The ASX-4000 also offers a mixed 30 Gbps configuration providing full redundancy for half of
the switch (see Figure G.6). In this mode, two fabrics each provide 10 Gbps non-redundant
capacity while the other two fabrics are configured as a single 10 Gbps working/protection pair.
This configuration allows network providers to offer high value, loss sensitive services on half of
the switch, while preserving 20 Gbps of bandwidth for lower priority data traffic. Additionally,
because each half of the switch can be configured independently of the other, the ASX-4000 eases
network migration issues. For example, a service provider could roll out non-redundant data
services using half of the ASX-4000's capacity, and later deploy higher value redundant service
on the same platform without disrupting the already existing customer base.

Fabric 1
40 Gbps
5 Gbps Port Card1CD
40 Gbps
Backplane
40 Gbps Fabric 2
5 Gbps Port Card 2CD Working
Protection
5 Gbps Port Card 3AB Non-redundant

40 Gbps Fabric 3

40 Gbps Fabric 4
Figure G.6 - ASX-4000 10 Gbps Redundant and 20 Gbps Non-redundant Configuration
G - 10 ATM Switch Network Configuration Manual

ASX-4000 Redundancy
If a switch half is being configured as redundant

NOTE and contains fabrics with revisions prior to E, the
operation will not be allowed. Fabrics prior to
Rev E will not be recognized, initialized, or
function if present during any attempt to
perform redundancy-related operations.
G.3.3 10 Gbps Fully Redundant Configuration

Figure G.7 shows a 10 Gbps fully redundant configuration. If fabric 1 is the working fabric,
ASX-4000 Redundancy
then fabric 2 is the protection fabric. Alternatively, you can choose to define fabric 2 as the
working fabric and fabric 1 as the protection fabric.

Fabric 1
40 Gbps
5 Gbps Port Card1CD
40 Gbps
Backplane
40 Gbps Fabric 2
Working
Protection
40 Gbps
40 Gbps
Figure G.7 - ASX-4000 10 Gbps Redundant Configuration
ATM Switch Network Configuration Manual G - 11

ASX-4000 Redundancy
G.3.4 Line Level and Port Card Redundancy Without Fabric

Redundancy
It is possible to have line level and port card redundancy without fabric redundancy by con-
figuring the switch for redundant mode without physically installing the redundant fabric
(see Figure G.8). However, as soon as the missing fabric is installed, it will initialize as a pro-
tection fabric and now provide fabric redundancy. In redundant mode there will be no impact
to capacity whether a redundant fabric is present or not.

Fabric 1
40 Gbps
5 Gbps Port Card1CD
40 Gbps
Backplane
40 Gbps
Working
Protection
40 Gbps
40 Gbps
Figure G.8 - ASX-4000 Line Level and Port Card Redundancy without Fabric Redundancy

ASX-4000 Redundancy
G.3.5 Fabric Redundancy Without Line Level and Port Card

Redundancy
It is also possible to achieve fabric redundancy without line level and port card redundancy
(see Figure G.9). This can be done by configuring the switch for redundant mode without
installing the redundant port cards. As soon as the missing port cards are installed, they will
initialize as a protection port card and thus provide line level and port card redundancy. In
redundant mode there will be no impact to capacity whether the redundant port cards are
present or not.
ASX-4000 Redundancy
Fabric 1
40 Gbps
5 Gbps Port Card1CD
40 Gbps
Backplane
5 Gbps
40 Gbps Fabric 2
5 Gbps
Working
Protection
40 Gbps
40 Gbps
Figure G.9 - ASX-4000 Fabric Redundancy without Line Level and Port Card Redundancy

ASX-4000 Redundancy
G.4 Redundant Link Naming Convention

To indicate that a link is configured for redundancy, an “R” is added to the front of the port
name. The following table shows the redundancy port naming for the working and protection
ports configured for redundancy:
Working port in a non- Protection port in a Port name in a redundant

redundant configuration redundant configuration configuration
1A1 2A1 R1A1
2A1 1A1 R2A1
1A1[3] 2A1[3] R1A1[3]
2A1[3] 1A1[3] R2A1[3]
For example, in AMI redundant links would display as follows:
myswitch:connections path term-> show

R1A1 0 term N/A 0.8K 1 6000 5002 pvc
R1A1 0 orig N/A 0.8K 1 6000 5002 pvc
R1A2 0 term N/A 0.8K 1 6000 4953 pvc
R1A2 0 orig N/A 0.8K 1 6000 4953 pvc
R1A3 0 term N/A 0.8K 1 511 4 pvc
R1A3 0 orig N/A 0.8K 1 511 4 pvc
R1A4 0 term N/A 0.8K 1 511 4 pvc
R1A4 0 orig N/A 0.8K 1 511 4 pvc
R1B1 0 term N/A 0.8K 1 6000 5002 pvc
R1B1 0 orig N/A 0.8K 1 6000 5002 pvc
R1B2 0 term N/A 0.8K 1 511 4 pvc
R1B2 0 orig N/A 0.8K 1 511 4 pvc
R1B3 0 term N/A 0.8K 1 511 4 pvc
R1B3 0 orig N/A 0.8K 1 511 4 pvc
R1B4 0 term N/A 356.2K 1 511 5 pvc
R1B4 0 orig N/A 356.2K 1 511 5 pvc
R1C1 0 term N/A 0.8K 1 6000 5002 pvc
R1C1 0 orig N/A 0.8K 1 6000 5002 pvc
R1C2 0 term N/A 356.2K 1 511 5 pvc
R1C2 0 orig N/A 356.2K 1 511 5 pvc
R1C3 0 term N/A 356.2K 1 511 5 pvc
R1C3 0 orig N/A 356.2K 1 511 5 pvc

ASX-4000 Redundancy
G.5 Configuring Working and Protection Interfaces

The ASX-4000 can only be configured for redundant fabric and port card operations with
Revision E fabrics. Fabric revisions prior to E do not support the redundancy feature. Revision
E fabrics are not supported in any release prior to ForeThought 6.1.
For redundancy, the ASX-4000 can be configured as a fully redundant 20 Gbps switch with full
1+1 data path redundancy (ports, port cards, and fabrics are protected on a 1+1 basis) or a
mixed 30 Gbps configuration providing full redundancy for half of the switch. In this configu-
ration, two fabrics each provide 10 Gbps non-redundant capacity while the other two fabrics
ASX-4000 Redundancy
are configured as a single 10 Gbps working/protection pair.
If an ASX-4000 has already been configured and

NOTE is operating in redundant mode (with two
Revision E fabrics) and an earlier revision fabric
is then hot swapped in, it will not be recognized
and consequently not initialized. The ASX-4000
will at this point only have one valid fabric, and
effectively no fabric redundancy for that switch
half.
An ASX-4000 containing earlier revision fabrics

NOTE having its CDB restored from a previously
configured valid redundant mode (that
originally contained Revision E fabrics) will fail
to initialize as redundant. This previous
redundant switch half will not be operational
until it is reconfigured as non-redundant.
APS is not supported if a pair of ports connected

NOTE from a non-redundant FORE switch is connected
to a working/protection port pair on a
redundant ASX-4000.

ASX-4000 Redundancy
Do not connect a protected port pair with a

NOTE remote switch that is not configured for
redundancy. To ensure proper signalling
operation, a protected port-pair (1A1/2A1)
should only be connected to the protected port-
pair of a remote switch configured for
redundancy.
The protection fabric is fixed and is dependent on the fabric that is being protected (working).
The following table shows the working fabric and the corresponding partner protection fabric:
Working Fabric Protection Fabric

1 2
2 1
3 4
4 3
When configuring redundancy, it is

NOTE recommended to configure the working interface
first.
The following example shows how to configure full redundancy on an ASX-4000:
For a detailed explanation of the redundancy

NOTE AMI commands, see Part 2 of the AMI
For descriptions of LED behavior specific to

NOTE redundancy, see Chapter 1 of the ASX-4000
Installation and Maintenance Manual.

ASX-4000 Redundancy
For a port card that is configured in a redundant

NOTE configuration between two ASX-4000 switches, if
a port card is disabled (insertion/ejection lever
in the unlocked position), the port card must be
pulled out of the slot. If a port card is disabled on
the local switch and experiences a power cycle,
when the switch reboots, it will turn on the
optical TX on the disabled ports. The remote
switch will show a valid carrier state and may
select this path as the active port. If this port is
ASX-4000 Redundancy
active on the remote switch it will use this port to
listen for UNI and SPANS signalling
information. The local port is still disabled and is
not transmitting signalling traffic. No signalling
will be established.
1. Specify the fabric(s) that you want protected (working):
redundancy fabric protect -fabric 1
Partner fabric 2 will be the protection fabric and fabric 1 will be the working fabric.
redundancy fabric protect -fabric 3
Part fabric 4 will be the protection fabric and fabric 3 will be the working fabric.
Fabric and port card compatibility is checked during the execution of this com-
mand. The protect command fails if any fabric is incompatible (checked for
Revision E fabrics), any port card is incompatible with its partner port card or if the
fabric which is being protected or unprotected is not present in the switch.
At this point, these commands are in pending mode. Any pending action takes
effect only after executing the commit command.
2. Execute the pending command(s):
redundancy fabric commit
The following displays:
WARNING: This command will delete the configuration values corresponding

to the protection fabric(s) and all the netmods connected to the
protection fabric. This command will also reboot the switch
Do you wish to continue? (y or n):

ASX-4000 Redundancy
When prompted, enter y to proceed with the pending protect command.

Fabric and port card compatibility is checked during the execution of this com-
mand. The commit command fails if any fabric is incompatible (checked for
Revision E fabrics), any port card is incompatible with its partner port card or if the
fabric which is being protected or unprotected is not present in the switch.
Executing the commit command deletes the configuration values corresponding
to the protection fabric(s) and all interface groups connected to the protection fab-
ric(s). Proceeding with this command will reboot the switch.
After confirmation to proceed with the commit command, all configuration infor-
mation for any connections originating or terminating on the portcards connected
to the primary bus of the protection fabric will be deleted. When executing a pend-
ing unprotect command, the protection fabric and its associated port card are
initialized in the same manner, as they would be following a CDB reset.
3. After the switch has rebooted, verify that the interfaces have been configured
correctly:
redundancy fabric show
Fabric Admin Mode Oper State Cloning State Pending Mode Committed
1 working active done none yes
2 protection standby done none yes
3 working active done none yes
4 protection standby done none yes

ASX-4000 Redundancy
4. Display APS group information:
redundancy port apsgroup show
Group Name Aps Command Line State Aps

Working Protection Mode
aps1A1 no_request none none uni
aps1A2 no_request none none uni
aps1A3 no_request sf sf uni
aps1A4 no_request sf sf uni
ASX-4000 Redundancy
aps1B1 no_request sf none uni
aps1B2 no_request none none uni
aps1C1 no_request none sf uni
aps1D1 no_request none sf uni
5. Modify the APS command for a particular port group:
redundancy port apsgroup modify aps1A1 -command forceswitch_working
For detailed information about the APS

NOTE commands, see Section 3.2.1.1 in Part 2 of the

ASX-4000 Redundancy
G.6 Partner Synchronization

Any modifications made to either the working/protection SONET ports/port cards/fabrics in
an ASX-4000 will also be applied to its partner component if present. The following sections
detail the commands that are applied to both working and protection links.
G.6.1 SONET Port Level Commands

Any modifications applied to a SONET port that is part of a redundant group will also be
applied to the partner SONET port.
myswitch:hardware port sonet-> modify

Usage:
[-hwportname] <BNP> Name
[[-mode] (sonet|sdh)] Mode
[[-loopback] <loopback>] Loopback Config
[[-timing] (network|internal)] Tx Clock Source
[[-scrambling] (on|off)] Scrambling
[[-emptycells] (unassigned|idle)] Empty Cells
[[-lineBipThrSeconds] <integer>] Line BIP Defect Seconds Threshold
[[-lineBipThrErrors] <integer>] Line BIP Defect Errors Threshold
[[-lineBipFailEnable] (enabled|disabled)] Fail on Excessive Errors
For example, if fabric 1 is the working fabric and fabric 2 is the protection fabric, modifying a
parameter for one fabric will modify the parameter for the partner fabric.
Entering the following:
hardware port sonet modify -hwportname 1A1 -scrambling off
OR
hardware port sonet modify -hwportname 2A1 -scrambling off
will turn scrambling off for both 1A1 and 2A1 if configured for redundancy.

ASX-4000 Redundancy
G.6.2 Netmod Level Commands

Any modifications applied to a port card that is attached to a fabric configured for redun-
dancy will also be applied to the partner port card.
G.6.2.1 Series 1 Port Cards

For Series 1 port cards, modifying the port card parameters for a Series 1 port card attached to
a redundant fabric will also modify the Series 1 partner port card.
ASX-4000 Redundancy
myswitch:hardware netmod traffic pc1-> modify
Usage:
[-netmod] <Netmod> Netmod
[[-model] <integer>] Model used
[[-epd] <percent>] Early Packet Discard
[[-epdubr] <percent>] Early Packet Discard for UBR
[[-efcion] <integer>] EFCI On
[[-efcioff] <integer>] EFCI Off
[[-vbrpriority] (rt|nrt)] VBR Priority (LE only)
parameter for one Series 1 port card will modify the parameter for the partner port card.
hardware netmod traffic pc1 modify -module 1A -epd 90
OR
hardware netmod traffic pc1 modify -module 2A -epd 90
will set the AAL5 packet drop value to 90 for both 1A and 2A if configured for redundancy.

ASX-4000 Redundancy
G.6.2.2 Series 2 Port Cards

For Series 2 port cards, modifying the port card parameters for a Series 2 port card attached to
a redundant fabric will also modify the Series 2 partner port card.
myswitch:hardware netmod traffic pc2-> modify

Usage:
[-netmod] <netmod> Netmod
[[-ucastConns] <integer>] Unicast Connections
[[-efcion] <integer>] EFCI On
[[-efcioff] <integer>] EFCI Off
[[-aal5cntMode] (cell|packet)] AAL5 Counting Mode
[[-aal5cntOverride] (on|off)] AAL5 Counting Mode Override
parameter for one Series 2 port card will modify the parameter for the partner port card.
hardware netmod traffic pc2 modify -module 1A -efcion 90
OR
hardware netmod traffic pc2 modify -module 2A -efcion 90
will set the EFCI threshold value to 90 for both 1A and 2A if configured for redundancy.

ASX-4000 Redundancy
G.6.2.3 Buffer Class and Buffer Class Assignment

Creating a new buffer class for a port card or assigning a buffer class to a service category will
create the same buffer class and buffer class assignment for both the port card and its partner
port card.
For example, if fabric 1 is the working fabric and fabric 2 is the protection fabric, creating a
new buffer class and assignment port card will also modify the parameter for the partner port
card.
ASX-4000 Redundancy
hardware netmod traffic pc2 buffer new -netmod 1A -buffer ABR
OR
hardware netmod traffic pc2 buffer new -netmod 2A -buffer ABR
will create a buffer class in both 1A and 2A if configured for redundancy.

hardware netmod traffic pc2 buffer assign new -netmod 1A -buffer ABR
OR
hardware netmod traffic pc2 buffer assign new -netmod 2A -servcat CBR
-subcat 1 -buffer CBR
will create a buffer class assignment in both 1A and 2A if configured for redundancy.

ASX-4000 Redundancy
G.6.3 Fabric Level Commands

Any modifications applied to a fabric that is configured for redundancy will also be applied to
the partner fabric.
myswitch:hardware fabric-> modify

Usage:
[-fabric] <board> Fabric
[[-mcastspace] <mcastSpace>] Multicast Space
myswitch:hardware fabric traffic-> modify

Usage:
[-fabric] <board> Fabric
[-priority] <fabric priority> Priority
[-feature] <traffic_feature> Feature
[[-value] <percent>] Value
fabric will also modify the partner fabric.
hardware fabric modify -fabric 1 -mcastspace mcast512
OR
hardware fabric modify -fabric 2 -mcastspace mcast512
will modify the number of multicast connections for both fabric 1 and fabric 2.
hardware fabric traffic modify -fabric 1 -priority 1 -feature 1 -value

80
OR
hardware fabric traffic modify -fabric 2 -priority 1 -feature 1 -value

80
will set the value to 80 for both fabric 1 and fabric 2 if configured for redundancy.

ASX-4000 Redundancy
G.6.4 Channelized OC-12 Port Card Commands

Any modifications applied to a channelized OC-12 port card that is attached to a fabric config-
ured for redundancy will also be applied to the partner port card.
G.6.4.1 SONET Medium

Any modifications applied to the admin status of a SONET medium is synchronized to the
partner SONET medium.
ASX-4000 Redundancy
myswitch:interfaces sonet medium-> modify
Usage:
[-ifindex] <SonetLowLevelIf> Sonet Interface
[[-type] <Operating Mode>] Framing Type
[[-loopbackmode] <sonet_loopconfig>] Loopback Config
[[-txclocksource] <sonet_clock>] Transmit Clock Source
[[-adminstatus] <ifadminstatus>] Admin Status
For example, if fabric 1 is the working fabric and fabric 2 is the protection fabric, modifying
the admin status of a SONET medium will also modify the partner SONET medium.
interfaces sonet medium modify -ifindex 1A -adminstatus down
OR
interfaces sonet medium modify -ifindex 2A -adminstatus down
will set the admin status to down for both 1A1 and 2A1 if configured for redundancy.

ASX-4000 Redundancy
G.6.4.2 SONET Path

Any modifications applied to the admin status of a SONET path is synchronized to the part-
ner SONET path.
myswitch:interfaces sonet path near-end-> modify

Usage:
[-ifindex] <SonetPathIf> If Index
[[-width] <sonet_pathwidth>] Width
[[-rdi] <sonet_rdi>] RDI Generation
[[-txsignallabel] <integer>] Tx Signal Label
[[-configfebes] <sonet_pathfebe>] FEBE Generation
[[-loopbackmode] <sonet_path_loopconfig>] Loopback Config
the admin status of a SONET path will also modify the partner SONET path.
interfaces sonet path near-end modify -ifindex 1A1[1,1] -adminstatus

down
OR
interfaces sonet path near-end modify -ifindex 2A1[1,1] -adminstatus

down
will set the admin status to down for both 1A1[1,1] and 2A1[1,1] if configured for redundancy.

ASX-4000 Redundancy
G.6.4.3 DSX3 Interface

Any modifications applied to the admin status of a DSX3 interface is synchronized to the part-
ner DSX3 interface.
myswitch:interfaces dsx3 near-end-> modify

Usage:
[-dsx3lineindex] <Dsx3If> Dsx3 Interface
[[-linelength] <DSX3 Line Length>] Line Length
[[-linetype] <DSX3 Application>] Line Type
ASX-4000 Redundancy
[[-linecoding] <DSX3 Line Coding>] Line Coding
[[-sendcode] <DSX3 FEAC Code>] Send Code
[[-circuitidentifier] <text>] Tx Circuit Identifier
[[-loopbackconfig] <DSX3 Loopback Configuration>]
Loopback Config
[[-transmitclocksource] <DSX3 Transmit Clock Source>]
Transmit Clock Source
the admin status of a DSX3 interface will also modify the partner DSX3 interface.
interfaces dsx3 near-end modify -dsx1lineindex 1A1[1,1] -adminstatus

down
OR
interfaces dsx3 near-end modify -dsx1lineindex 2A1[1,1] -adminstatus

down
will set the admin status to down for both 1A1[1,1] and 2A1[1,1] if configured for redundancy.

ASX-4000 Redundancy
G.6.4.4 ATM Transmission Convergence

Any modifications applied to the ATM-TC is synchronized to the partner ATM-TC.
myswitch:interfaces atm-tc-> modify

Usage:
[-ifindex] <TcIf> If Index
[[-cellscrambling] <cellscrambling>] Cell Payload Scrambling
[[-emptycell] <emptycell>] Empty Cell
[[-loopback] <loopback>] Loopback
[[-framingmode] <framingmode>] Framing Mode
the framing mode for this ATM-TC will also modify the partner ATM-TC.
interfaces atm-tc modify -ifindex 1A1[1,1] -framingmode plcp
OR
interfaces atm-tc modify -ifindex 2A1[1,1] -framingmode plcp
will set the framing mode to plcp for both 1A1[1,1] and 2A1[1,1] if configured for redundancy.

ASX-4000 Redundancy
G.6.5 LED Model

The setting of the LED model is synchronized to the partner port for those ports that are con-
figured for redundancy.
myswitch:hardware port led-> modify
Usage:
[-hwportname] <BNP> Name
[[-ledmodel] (lan1|wan1|lan2|wan2)] LED Model
For example, if fabric 1 is the working fabric and fabric 2 is the protection fabric, setting the
ASX-4000 Redundancy
LED model for one port will also modify the partner port.
hardware port led modify -hwportname 1A1 -ledmodel lan1
OR
hardware port led modify -hwportname 2A1 -ledmodel lan1
will set the LED model to LAN1 for both 1A1 and 2A1 if configured for redundancy.
G.6.6 Netmod Alarms

The modification of alarm priorities of a port card slot is synchronized to the partner port card
slot.
myswitch:hardware netmod alarms-> modify
Usage:
[-slot] <Slot> Slot
[[-priority] <alarm_priority>] Priority
the alarm priorities for a port card slot will also modify the partner port card slot.
hardware netmod alarms modify -slot 1A -priority high
OR
hardware netmod alarms modify -slot 2A -priority high
will set the alarm priority to high for both slots 1A and 2A if configured for redundancy.

ASX-4000 Redundancy
G.7 Loopbacks on Redundant Ports

When configuring loopbacks for redundant ports, both, diagnostic, path, or other loop-
backs will be active only on the port on which the loopback was executed. Loopbacks will not
be active on the partner port. For example, if both, diagnostic, path, or other loopback is
executed on 1A1, entering redundancy port show will display the Oper Mode for 1A1 as
loopback_on and it’s partner port 2A1 would display its Oper Mode as loopback_off.
Line loopbacks apply to both working and protection ports. For example, if line loopback is
executed on 1A1, entering redundancy port show will display the Oper Mode for 1A1 as
loopback_on and its partner port 2A1 would also display its Oper Mode as loopback_on.
If you want to perform diagnostics on 1A1 and 1A1 and 2A1 are part of an APS redundant
group, you would do the following:
1. Configure the port for loopback by entering the following:
hardware port sonet modify 1A1 -loopback diag
The same configuration on channelized OC-12

NOTE port cards would use the interfaces sonet
medium modify <SonetLowLevelIf>
-loopbackmode command.
2. Display the operational state of the redundant ports by entering the following:
redundancy port show
PortName Admin Mode Oper Mode GroupName

1A1 working loopback_on aps1A1
1B1 working active aps1B1
2A1 protection loopback_off aps1A1
2B1 protection standby aps1B1
If line loopback was configured for 1A1, the

NOTE operational state of its partner port 2A1 would
display loopback_on also.

Index
Numerics automatic protection switching

1+1 SONET APS . . . . . . . . . . . . . . . . . . . . . .G - 1 (APS). . . . . . . . . . . . . .See 1+1 SONET APS
Available . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 29
A
Available Bit Rate (ABR) . . . . . . . . . . . . . . 1 - 38
ABR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 38
Available Cell Rate Minimum
active detection Threshold (AVCRMT) . . . . . . . . . . . . 6 - 29
basic connections . . . . . . . . . . . . . . . . B - 4 Available Cell Rate Proportional
CAS connections . . . . . . . . . . . . . . . . B - 3 Multiplier (AVCRPM) . . . . . . . . . . . . 6 - 29
address registration . . . . . . . . . . . . . . . . . . . 2 - 4 AVCRMT . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 29
address summarization . . . . . . . . . . . . . . . 6 - 15 AVCRPM . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 29
aggregation token . . . . . . . . . . . . . . . . . . . . 6 - 14
B
AMI access level
basic connections
network . . . . . . . . . . . . . . . . . . . . . . . 8 - 16
active detection . . . . . . . . . . . . . . . . . B - 4
serial . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 16
idle channel suppression . . . . . . . . . B - 3
AMI command privileges . . . . . . . . . . . . . 8 - 15
idle detection using mask patterns . B - 3
admin . . . . . . . . . . . . . . . . . . . . . . . . 8 - 15
idle detection using pattern
user . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 15
matching . . . . . . . . . . . . . . . . . B - 3
anycast address scoping . . . . . . . . . . . . . . 6 - 16
border node . . . . . . . . . . . . . . . 6 - 7, 6 - 10, 6 - 11
application
border node with LGN subset . . . . . . . . . 6 - 10
configuring on a FramePlus
Broadcast and Unknown Server (BUS) . . . 3 - 3
network module . . . . . . . . . .D - 22
broadcast packets . . . . . . . . . . . . . . . . . . . . . 4 - 5
area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 19
BT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 39
ARP cache . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 8
Burst Tolerance (BT) . . . . . . . . . . . . . . . . . 1 - 39
ARP reply . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 8
BUS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3
ARP request . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 8
ARP server . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 6 C
configuring a FORE switch as . . . . . 2 - 7 CAS connections
ATM ARP (ATM address resolution active detection . . . . . . . . . . . . . . . . . B - 3
protocol) . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 6 idle channel suppression . . . . . . . . . B - 2
ATM Forum LAN Emulation Over idle signalling pattern matching . . . B - 5
ATM Version 1.0 . . . . . . . . . . . . . . . . . . 4 - 1 CBR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 38
ATM Forum PNNI . . . . . . . . . . . . . . . . . . . . 6 - 1 CDVT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 39
ATMF border node . . . . . . . . . . . . . . . . . . 6 - 10 Cell Delay Variation Tolerance (CDVT) . 1 - 39
ATM Switch Network Configuration Manual Index - 1

Index
Cell Loss Priority (CLP) bit . . . . . . . . . . . . 1 - 40 defining an ELAN . . . . . . . . . . . . . . . . . . . 3 - 31

CES (Circuit Emulation Services) . . . . . . . B - 1 Designated Transit Lists (DTLs) . . . . . . . . 6 - 8
CES connection directed DTL . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 8
creating . . . . . . . . . . . . . . . . . . . . . . . . B - 9 Distributed LAN Emulation (DLE) . . 3 - 9, 4 - 6
deleting . . . . . . . . . . . . . . . . . . . . . . . . B - 8 distributed timing . . . . . . . . . . . . . . . . . . . . 9 - 1
displaying . . . . . . . . . . . . . . . . . . . . . B - 15 ASX-1000 . . . . . . . . . . . . . . . . . . . . . . 9 - 4
displaying advanced information B - 16 ASX-1200 . . . . . . . . . . . . . . . . . . . . . . 9 - 4
displaying idle channel ASX-200BX . . . . . . . . . . . . . . . . . . . . . 9 - 4
suppression information . . . B - 17 ASX-200WG . . . . . . . . . . . . . . . . . . . . 9 - 4
displaying statistics . . . . . . . . . . . . . B - 18 ASX-4000 . . . . . . . . . . . . . . . . . . . . . . 9 - 6
child node . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 13 ESX-3000 . . . . . . . . . . . . . . . . . . . . . . . 9 - 6
Circuit Emulation Services (CES) . . . . . . . B - 1 ForeRunnerLE 155 . . . . . . . . . . . . . . 9 - 5
Classical IP ATM network ForeRunnerLE 25 . . . . . . . . . . . . . . . 9 - 5
configuring . . . . . . . . . . . . . . . . . . . . . 2 - 1 TNX-1100 . . . . . . . . . . . . . . . . . . . . . . 9 - 4
Classical IP interface . . . . . . . . . . . . . . . . . . 2 - 2 TNX-210 . . . . . . . . . . . . . . . . . . . . . . . 9 - 4
Classical IP over ATM . . . . . . . . . . . . . . . . . 2 - 1 DLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 9
Classical IP PVC . . . . . . . . . . . . . . . . . . . . . 2 - 10 DLE peer servers
CLP bit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 40 starting . . . . . . . . . . . . . . . . . . . . . . . 3 - 41
configuration domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 21
emulated LAN, example . . . . . . . . . 4 - 3 DTL (Designated Transit List) . . . . . . . . . . 6 - 8
congestion-based UBR routing . . . . . . . . 6 - 27 dynamic path . . . . . . . . . . . . . . . . . . . . . . . 7 - 24
enabling . . . . . . . . . . . . . . . . . . . . . . 6 - 29
E
connection
egress rate enforcement
configuration-direct . . . . . . . . . . . . . 3 - 4
configuring . . . . . . . . . . . . . . . . . . . . D - 10
control-direct . . . . . . . . . . . . . . . . . . . 3 - 4
ELAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 1
control-distribute . . . . . . . . . . . . . . . 3 - 4
components . . . . . . . . . . . . . . . . . . . . 3 - 2
data-direct . . . . . . . . . . . . . . . . . . . . . 3 - 4
configuring . . . . . . . . . . . . . . . . 3 - 1, 3 - 22
multicast-forward . . . . . . . . . . . . . . . 3 - 4
joining . . . . . . . . . . . . . . . . . . . . . . . . 3 - 44
multicast-send . . . . . . . . . . . . . . . . . . 3 - 4
ELAN access control . . . . . . . . . . . . 3 - 21, 3 - 31
connection process, LEC . . . . . . . . . . . . . . . 4 - 5
disabling . . . . . . . . . . . . . . . . . . . . . . 3 - 21
connection scope selection . . . . . . . . . . . . 6 - 16
enabling . . . . . . . . . . . . . . . . . . . . . . 3 - 21
Constant Bit Rate (CBR) . . . . . . . . . . . . . . 1 - 38
Emulated LAN (ELAN) . . . . . . . . . . . . . . . 3 - 1
containing path . . . . . . . . . . . . . . . . 7 - 18, 7 - 24
address resolution . . . . . . . . . . . . . . . 3 - 7
crankback . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 8
data transfer . . . . . . . . . . . . . . . . . . . . 3 - 8
D initialization . . . . . . . . . . . . . . . . . . . . 3 - 6
data encryption . . . . . . . . . . . . . . . . .8 - 7, 8 - 11 operation . . . . . . . . . . . . . . . . . . . . . . 3 - 4
database exchange protocol . . . . . . . . . . . . 6 - 2 registration . . . . . . . . . . . . . . . . . . . . . 3 - 7
Index - 2 ATM Switch Network Configuration Manual

Index
emulated LAN (ELAN) configuration examples . . . . . . . . . . B - 4

example configuration . . . . . . . . . . . 4 - 3 idle pattern configuration example B - 6
running multiple . . . . . . . . . . . . . . . . 4 - 6 integration period configuration
entry border node . . . . . . . . . . . . . . . . . . . . 6 - 11 example . . . . . . . . . . . . . . . . . . B - 7
ESI (End System Identifier) . . . . . . . 2 - 4, 6 - 32 mask patterns configuration
exit border node . . . . . . . . . . . . . . . . . . . . . 6 - 11 example . . . . . . . . . . . . . . . . . . B - 6
idle detection using mask patterns . . . . . . B - 3
F idle detection using pattern matching . . . B - 3
flooding protocol . . . . . . . . . . . . . . . . . . . . . 6 - 3 Idle detection using signalling bits . . . . . . B - 2
FORE IP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 3 idle timeslot . . . . . . . . . . . . . . . . . . . . . . . . . . B - 2
ForeThought PNNI . . . . . . . . . . . . . . . . . . . 5 - 1 ILMI (Interim Local Management
backbone topology . . . . . . . . . . . . . . 5 - 9 Interface) . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 4
border switches . . . . . . . . . . . . . . . . . 5 - 8 IMA network module
hello indication messages . . . . . . . . . 5 - 8 upgrading software . . . . . . . . . . . . . F - 12
link metrics . . . . . . . . . . . . . . . . . . . . . 5 - 8 InARP reply . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 8
peer group summary node . . 5 - 8, 5 - 9 InARP request . . . . . . . . . . . . . . . . . . . . . . . . 2 - 8
physical network . . . . . . . . . . . . . . . . 5 - 5 ingress rate enforcement
topology database . . . . . . . . . . . . . . . 5 - 5 configuring . . . . . . . . . . . . . . . . . . . .D - 10
two-level hierarchy . . . . . . . . . . . . . . 5 - 6 initialization process, LEC . . . . . . . . . . . . . 4 - 4
Frame-based UNI (FUNI) . . . . . . . . . . . . . .D - 1 integration period . . . . . . . . . . . . . . . . . . . . B - 2
FramePlus network module Interim Local Management Interface
configuring an application . . . . . . .D - 22 (ILMI) . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 4
FUNI (Frame-based UNI) . . . . . . . . . . . . . .D - 1 inverse ARP (InARP) . . . . . . . . . . . . . . . . . . 2 - 8
Inverse Multiplexing for ATM (IMA) . . . . F - 1
G
IP filtering . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 17
gateway switch . . . . . . . . . . . . . . . . . . . . . . 6 - 17
authorized IP address table . . . . . . 8 - 17
GCRA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 39
creating an authorized address . . . 8 - 17
Generic Cell Rate Algorithm (GCRA) . . . 1 - 39
deleting an authorized address . . . 8 - 17
H displaying authorized addresses . 8 - 17
Hello protocol . . . . . . . . . . . . . . . . . . . . . . . . 6 - 2 information about the last
hierarchical routing packet dropped . . . . . . . . . . . 8 - 18
PNNI . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 4 statistics . . . . . . . . . . . . . . . . . . . . . . . 8 - 18
horizontal links . . . . . . . . . . . . . . . . . . 6 - 2, 6 - 5 IP filtering flags . . . . . . . . . . . . . . . . . . . . . 8 - 17
I
idle channel suppression . . . . . . . . . . . . . . . B - 2
basic connections . . . . . . . . . . . . . . . . B - 3
CAS connections . . . . . . . . . . . . . . . . B - 2

Index
K local authentication . . . . . . . . . . . . . . . . . . . 8-5

Kerberos Logical IP Subnet (LIS) . . . . . . . . . . . . . . . . 2-2
data encryption . . . . . . . . . . . . . . . . 8 - 11 logical link (loglink) . . . . . . . . . . . . . . . . . . . 5-2
installing the server software . . . . 8 - 12 login authentication . . . . . . . . . . . . . . . . . . . 8-5
server . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 11 local . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-5
SecurID . . . . . . . . . . . . . . . . . . . . . . . . 8-5
L
loglink (logical link) . . . . . . . . . . . . . . . . . . . 5-2
LAN Emulation Client (LEC) . . . . . . . . . . . 3 - 3
lop-sided hierarchy . . . . . . . . . . . . . . . . . . . 6-7
connection process . . . . . . . . . . . . . . 4 - 5
initialization process . . . . . . . . . . . . . 4 - 4 M
LAN Emulation Configuration Server Maximum Burst Size (MBS) . . . . . . . . . . . 1 - 38
(LECS) . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3 MBS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 38
LAN Emulation Server (LES) . . . . . . . . . . . 3 - 3 MCR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 39
LAN Emulation services . . . . . . . . . . . . . . . 3 - 2 MIB overview . . . . . . . . . . . . . . . . . . . . . . . . E - 2
starting . . . . . . . . . . . . . . . . . . . . . . . 3 - 39 Minimum Cell Rate (MCR) . . . . . . 1 - 39, 6 - 27
leaky bucket algorithm . . . . . . . . . . . . . . . 1 - 39 multicast packets . . . . . . . . . . . . . . . . . . . . . 4 - 5
LEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3
N
starting . . . . . . . . . . . . . . . . . . . . . . . 3 - 44
network configuration examples . . . . . . . 2 - 11
LECS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3
nodal aggregation . . . . . . . . . . . . . . . . . . . 6 - 15
well-known address . . . . . . . . . . . . 3 - 44
node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 5
LECS configuration file
node secret file . . . . . . . . . . . . . . . . . . . . . . . 8 - 8
configuring DLE . . . . . . . . . . . . . . . 3 - 31
NSAP addresses . . . . . . . . . . . . . . . . . . . . . . 2 - 4
configuring MPOA . . . . . . . . . . . . . 3 - 34
NSAP prefix . . . . . . . . . . . . . . . . . . . . 2 - 4, 6 - 32
ELAN access control . . . . . . . . . . . . 3 - 31
sample . . . . . . . . . . . . . . . . . . . . . . . . 3 - 36 O
starting . . . . . . . . . . . . . . . . . . . . . . . 3 - 39 organizational scope . . . . . . . . . . . . . . . . . 6 - 16
syntax . . . . . . . . . . . . . . . . . . . . . . . . 3 - 25 originating path . . . . . . . . . . . . . . . . . . . . . . 1 - 6
LECS control parameters . . . . . . . . . . . . . 3 - 33 originating paths . . . . . . . . . . . . . . . . . . . . . 1 - 6
LES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3 displaying . . . . . . . . . . . . . . . . . 1 - 9, 1 - 10
starting . . . . . . . . . . . . . . . . . . . . . . . 3 - 41 outside link . . . . . . . . . . . . . . . . . . . . . 6 - 7, 6 - 11
levels in PNNI . . . . . . . . . . . . . . . . . . . . . . 6 - 21
P
LGN (logical group node) subset . . . . . . 6 - 10
parent node . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 13
link aggregation . . . . . . . . . . . . . . . . . . . . . 6 - 15
PC congestion range . . . . . . . . . . . . . . . . . 6 - 29
Link Management Interface (LMI) . . . . . D - 11
PCR . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 38, 1 - 39
link-scope UNI . . . . . . . . . . . . . . . . . . . . . . 7 - 23
Peak Cell Rate (PCR) . . . . . . 1 - 38, 1 - 39, 6 - 27
LIS (Logical IP Subnet) . . . . . . . . . . . . . . . . 2 - 2
peer group leader (PGL) . . . . . . . . . . . . . . . 6 - 4
LMI (Link Management Interface) . . . . . D - 11
load-balanced routing . . . . . . . . . . . . . . . . 6 - 26

Index
permanent virtual channel redundancy. . . . . . . . . . . .See 1+1 SONET APS

displaying advanced information . 1 - 16 related manuals . . . . . . . . . . . . . . . . . . . . . . . . .ii
permanent virtual circuits (PVCs) . . . . . . . 1 - 1 RFC-1577 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 1
PGL RMON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E - 1
peer group leader . . . . . . . . . . . . . . . 6 - 5 RMON overview . . . . . . . . . . . . . . . . . . . . . E - 4
PGL (peer group leader) subset . . . . . . . . 6 - 10
S
PGL elections . . . . . . . . . . . . . . . . . . . . . . . 6 - 13
scope . . . . . . . . . . . . . . . . . . . . . . . . . .6 - 16, 6 - 24
PMP PNNI SPVC
SCR . . . . . . . . . . . . . . . . . . . . . . . . . . .1 - 38, 1 - 39
example . . . . . . . . . . . . . . . . . . . . . . . 1 - 35
sdconf.rec file . . . . . . . . . . . . . . . . . . . . . . . . 8 - 7
PNNI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 1
editing . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 8
gateway switch . . . . . . . . . . . . . . . . 6 - 17
transferring to the switch . . . . . . . . . 8 - 7
split switch . . . . . . . . . . . . . . . . . . . . 6 - 17
SecurID
PNNI path computation . . . . . . . . . . . . . . . 6 - 3
data encryption . . . . . . . . . . . . . . . . . 8 - 7
PNNI peer group . . . . . . . . . . . . . . . . . . . . . 6 - 5
node secret file . . . . . . . . . . . . . . . . . . 8 - 7
PNNI policies . . . . . . . . . . . . . . . . . . . . . . . 6 - 23
passcode . . . . . . . . . . . . . . . . . . . . . . . 8 - 6
PNNI profile
personal identification number
congestion-based . . . . . . . . . . . . . . . 6 - 30 (PIN) . . . . . . . . . . . . . . . . .8 - 5, 8 - 6
PNNI signalling protocol . . . . . . . . . . . . . . 6 - 8 PIN number . . . . . . . . . . . . . . . . . . . . 8 - 6
PNNI SPVC server . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 6
configuring for Frame Relay . . . . .D - 20 tokens . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 6
configuring for FUNI . . . . . . . . . . .D - 29 securid
PNNI Topology State Elements . . . . . . . . . 6 - 3 installing the server software . . . . . 8 - 7
PNNI Topology State Packets . . . . . . . . . . 6 - 3 SecurID authentication . . . . . . . . . . . . . . . . 8 - 5
point-to-point PNNI SPVC security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 1
creating . . . . . . . . . . . . . . . . . . . . . . . 1 - 33 selector field . . . . . . . . . . . . . . . . . . . . . . . . . 2 - 4
policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 23 Smart Permanent Virtual Circuit . . . . . . . 1 - 30
advertise . . . . . . . . . . . . . . . . . . . . . . 6 - 23 SNMP overview . . . . . . . . . . . . . . . . . . . . . . E - 3
summary . . . . . . . . . . . . . . . . . . . . . . 6 - 23 SNMP trap . . . . . . . . . . . . . . . . . . . . . . . . . . .A - 2
suppress . . . . . . . . . . . . . . . . . . . . . . 6 - 23 adding . . . . . . . . . . . . . . . . . . . . . . . .A - 24
PTSEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 3 deleting . . . . . . . . . . . . . . . . . . . . . . .A - 25
PTSPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 3 displaying . . . . . . . . . . . . . . . . . . . . .A - 24
PVC revalidation . . . . . . . . . . . . . . . . . . . . 2 - 10 supported on the switch . . . . . . . . . .A - 2
R source area ID . . . . . . . . . . . . . . . . . . . . . . . 6 - 24
rate enforcement . . . . . . . . . . . . . . . . . . . . .D - 10 SPANS interface . . . . . . . . . . . . . . . . . . . . . . 2 - 3
configuring egress . . . . . . . . . . . . . .D - 10
configuring ingress . . . . . . . . . . . . .D - 10
reachability information . . . . . . . . . . . . . . . 6 - 2

Index
SPANS SPVC TNX-1100 . . . . . . . . . . . . . . . . . . . . . . 9 - 4

configuring . . . . . . . . . . . . . . . . . . . . 1 - 30 TNX-210 . . . . . . . . . . . . . . . . . . . . . . . 9 - 4
configuring for Frame Relay . . . . . D - 17 timing on a switch . . . . . . . . . . . . . . . . . . . . 9 - 1
configuring for FUNI . . . . . . . . . . . D - 26 traffic policing . . . . . . . . . . . . . . . . . . . . . . 1 - 39
split switch . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 17
U
SPVC
UBR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 38
configuring for Frame Relay . . . . . D - 17
Unspecified Bit Rate (UBR) . . . . . . . . . . . 1 - 38
configuring for FUNI . . . . . . . . . . . D - 26
UPC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 39
SPVC (Smart Permanent Virtual Circuit) 1 - 30
UPC contract
SRTS (Synchronous Residual Time
creating . . . . . . . . . . . . . . . . . . . . . . . 1 - 42
Stamp) . . . . . . . . . . . . . B - 11, B - 12, B - 13
UPC traffic contract parameters . . . . . . . 1 - 41
Sustainable Cell Rate (SCR) . . . . . . 1 - 38, 1 - 39
upgrading
SVCC RCC . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 12
IMA network module software . . F - 12
switchclock . . . . . . . . . . . . . . . . . . . . . . . . . . 9 - 1
uplink . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 11
failover . . . . . . . . . . . . . . . . . . . . . . . . 9 - 2
upnode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 11
Synchronous Residual Time Stamp
(SRTS) . . . . . . . . . . . . . . B - 11, B - 12, B - 13 Usage Parameter Control (UPC) . . . . . . . 1 - 39
userid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 1
T AMI access level . . . . . . . . . . . . . . . 8 - 16
tagging cells . . . . . . . . . . . . . . . . . . . . . . . . 1 - 40 AMI command privileges . . . . . . . 8 - 15
Technical Assistance Center (TAC) changing a password . . . . . . . . . . . 8 - 16
contacting . . . . . . . . . . . . . . . . . . . . . . . . iii configuring . . . . . . . . . . . . . . . . . . . . . 8 - 1
terminating path . . . . . . . . . . . . . . . . . . . . . 1 - 6 setting a password . . . . . . . . . . . . . 8 - 16
terminating paths . . . . . . . . . . . . . . . . . . . . . 1 - 6
displaying . . . . . . . . . . . . . . . . .1 - 9, 1 - 10 V
through paths . . . . . . . . . . . . . . . . . . . . . . . . 1 - 4 v5srvtab file . . . . . . . . . . . . . . . . . . . . . . . . 8 - 12
displaying advanced information . 1 - 8 Variable Bit Rate (VBR) . . . . . . . . . . . . . . . 1 - 38
timeslots . . . . . . . . . . . . . . . . . . . . . . . . . . . . B - 2 VBR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 38
timing VCI allocation range . . . . . . . . . . . . . . . . . 7 - 18
ASX-1000 . . . . . . . . . . . . . . . . . . . . . . 9 - 4 VC-space . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 23
ASX-1200 . . . . . . . . . . . . . . . . . . . . . . 9 - 4 virtual channel . . . . . . . . . . . . . . . . . . . . . . 1 - 11
ASX-200BX . . . . . . . . . . . . . . . . . . . . . 9 - 4 virtual channel identifier (VCI) . . . . . . . . . 1 - 1
ASX-200WG . . . . . . . . . . . . . . . . . . . . 9 - 4 virtual path . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 3
ASX-4000 . . . . . . . . . . . . . . . . . . . . . . 9 - 6 displaying advanced information . 1 - 8
ESX-3000 . . . . . . . . . . . . . . . . . . . . . . . 9 - 6 virtual path identifier (VPI) . . . . . . . . . . . . 1 - 1
ForeRunnerLE 155 . . . . . . . . . . . . . . 9 - 5 virtual path terminator
ForeRunnerLE 25 . . . . . . . . . . . . . . . 9 - 5 displaying advanced information 1 - 10
port level . . . . . . . . . . . . . . . . . . . . . . . 9 - 3

Index
virtual path terminators (VPTs) . . . . . . . . . 1 - 6

VP-scope UNI . . . . . . . . . . . . . . . . . . . . . . . 7 - 23

Index

ATM Switch Network Configuration Manual

Uploaded by

Copyright:

Available Formats

ATM Switch Network Configuration Manual

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ATM Switch Network Configuration Manual

Uploaded by

Copyright:

Available Formats

What are the main components of an ATM network?

What are the main components of an ATM network?

What are some of the configuration options discussed in the manual?

What are some of the configuration options discussed in the manual?

ATM Switch Network

Software Version 6.1.x

FCC CLASS A NOTICE

DOC CLASS A NOTICE

VCCI CLASS A NOTICE

FCC REQUIREMENTS (Notice to Users of DS1 Service)

Port ID REN/SOC FIC USOC

CANADIAN IC CS-03 COMPLIANCE STATEMENT

Pan European Approval - Marking

Network Module Connects to Approval Number

Required User Guide Statements - UK Installation

NM-8/E1D and NM-4/E1D

Ports Safety Status

Ports Safety Status

ALPHA/BETA TEST DISCLAIMER

CHAPTER 1 Configuring PVCs

ATM Switch Network Configuration Manual TOC - 1

1.5.4 Creating SPVCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 - 30

CHAPTER 2 Configuring Classical IP

CHAPTER 3 Configuring an Emulated LAN

TOC - 2 ATM Switch Network Configuration Manual

3.2.3 LAN Emulation Server (LES) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 3

ATM Switch Network Configuration Manual TOC - 3

3.7.10 Update the LECS.CFG File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 - 55

CHAPTER 5 ForeThought PNNI

TOC - 4 ATM Switch Network Configuration Manual

CHAPTER 6 ATM Forum PNNI

ATM Switch Network Configuration Manual TOC - 5

6.3.9.3 Link Aggregation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 - 15

TOC - 6 ATM Switch Network Configuration Manual

6.8.1.3.1 Enterprise Example . . . . . . . . . . . . . . . . . . . . . . . . 6 - 33

ATM Switch Network Configuration Manual TOC - 7

7.7.1 PNNI 1.0/UNI 4.0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 - 35

TOC - 8 ATM Switch Network Configuration Manual

8.1.2 AMI Command Privileges. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 - 15

CHAPTER 9 Configuring Timing

APPENDIX A Configuring SNMP

ATM Switch Network Configuration Manual TOC - 9

APPENDIX B Configuring Circuit Emulation Services

APPENDIX C Converting FT-PNNI and PNNI Networks

TOC - 10 ATM Switch Network Configuration Manual

C.3.1.1.1.3 Convert the Individual Peer

APPENDIX D Configuring FramePlus Modules

ATM Switch Network Configuration Manual TOC - 11

D.4.4.1 Creating a Frame Relay SPANS SPVC . . . . . . . . . . . . . . . . . D - 17

APPENDIX F Configuring IMA Modules

TOC - 12 ATM Switch Network Configuration Manual

F.4 Upgrading the IMA Network Module Software . . . . . . . . . . . . . . . . . . . . . . . . . . . F - 12

APPENDIX G ASX-4000 Redundancy

ATM Switch Network Configuration Manual TOC - 13

TOC - 14 ATM Switch Network Configuration Manual

CHAPTER 1 Configuring PVCs

CHAPTER 2 Configuring Classical IP

CHAPTER 3 Configuring an Emulated LAN

ATM Switch Network Configuration Manual LOF - 1

Figure 3.6 Distributed LAN Emulation Model . . . . . . . . . . . . . . . . . . . . . . . 3 - 12