Networking Key word Synchronous Transfer Mode (STM) Time-Division-Multiplexing (TDM) Circuit switching Routing: Routing Connection Oriented

Asynchronous Transfer Mode (ATM) Statistical Multiplexing (SM) Packets Switching Routing: Connection/Connectionless Oriented

PUBLIC SWITCHED TELEPHONE NETWORKS PSTN

Time Division Multiplexing

Assume that we have m communication terminals, T1, T2, .., Tm sharing a transmission line, how do we schedule the sharing of communication bandwidth? T1 T2 Tm

Assume that the bandwidth is shared by the terminals transmitting at different times.

We also assume that a scheduling mechanism is available so that the transmissions are conflict free, namely, that no two terminals attempt to transmit at the same time. We call this scheduled or arbitrated access communication.

BROADBAND BUS

SCHEDULER Multiplexing with scheduling

In the absence of an arbitration mechanism, two communication terminals may transmit at the same time, often resulting in unintelligible transmissions.

Two basic approaches to multiplexing:

1. The first approach assumes a common time reference among the terminals. We call this time reference a frame reference. The communication bandwidth assigned for each terminal is termed a circuit. This mode of multiplexing is commonly known as the Synchronous Transfer Mode (STM). 2. The second approach assumes no frame reference among the terminals, hence the name Asynchronous Transfer Mode (ATM). This mode allows more flexible sharing of bandwidth by avoiding rigid bandwidth assignments. Bandwidth is seized on demand, and the information transmitted (together with a proper label) upon a successful seizure is termed a packet.

The Asynchronous Transfer Mode

The definition of a frame depends on the bit-rates of the terminals multiplexed on the transmission link. The choice of frame structure is difficult since we have little knowledge of the traffic mix. An alternative approach abandons the concept of a frame reference altogether. Instead of choosing a basic terminal bit-rate as in TDM, ATM achieves more flexible bandwidth sharing allowing the terminals to seize bandwidth when a sufficient number of bits are generated. Without a frame reference, these bits have no implicit ownership, unlike STM for which each slot is assigned an owner. Hence a key feature of ATM is that information from each terminal must be labeled.

The Asynchronous Transfer Mode

There are many forms of asynchronous multiplexing:
First, we may have fixed length blocks of information from each terminal. These blocks are termed cells in ATM terminology. A cell is labeled block of transmitted information, and usually has a small information payload (typically from 32 bytes to 128 bytes). We shall also refer to them as short fixed length packets.

The Asynchronous Transfer Mode

Cell (or Short fixed length packets):
Each cell or packet has a fixed size of l bits. The channel is slotted into fixed intervals of duration l/C, each transporting a cell. The terminals are asynchronous in the sense that they have no common time reference other than the common slot reference. A label for each time slot must be provided by the terminal which transmits in that time slot.

The Asynchronous Transfer Mode

The label identifies the terminal generating the bits delivered in the time slot. A label is included in the header part of a packet. The header may serve other functions; such as classifying the information payload (type and priority), and possible error check sums for protecting the header from transmission error.
PACKET INFO

The Asynchronous Transfer Mode

There are two major factors in determining the proper packet size:
First, headers use up part of the communication capacity of the link. This overhead is inversely proportional to the packet size l, consequently favoring long packet. Second, a packetization delay is needed for the terminal to collect the l bits for a packet. The delay between signal generation and reception is given by , t = l/b plus the delay taken for the signal to travel in the network. For some applications, excessive delay results in perceivable degradation of the quality of communication. Consequently, minimizing packetization delay requires choosing short packets. A compromise has to be chosen between two opposing factors.

HEADER

t
l BITS SLOTS Multiplexing of Fixed Length Packets

The Asynchronous Transfer Mode

Variable Length Packets:
Instead of short fixed length packets, it is often convenient (particularly for data communications) to use long (say 128 bytes or more) variable length packets. Besides the label for ownership, the packet header should also contain the information for packet length to mark the end of the packet, as well as a flag to mark the beginning of the packet.
PACKET INFO

HEADER

t
l BITS SLOTS Multiplexing of Fixed Length Packets

PSTN
Local Exchange Carriers (LECs) LECs provide local telephone service, usually within the boundaries of a metropolitan area, state, or province. LECs also provide short-haul, long distance service, Centrex, certain enhanced services such as voice mail, and various data services. BOCS (Bell Operating Companies), originally were wholly owned by AT&T, dominated the ILECs landscape. Local Access and Transport Area (LATA) Effective January 1, 1984, those 22 BOCs were spun off from AT&T as a result of the Modified Final Judgement (MFJ). BOCs were reorganized into seven Regional BOCS (RBOCS). BOCs were limited to providing basic voice and data services within defined geographical areas, known as Local Access and Transport Areas (LATAs).

Are some 170 areas defined by the MFJ Collectively span all BOC territories

In general, each Boc territory comprises several LATAs

PSTN Continue InterExchange Carriers (IXCs or IECs) IXCs are responsible for long-haul, long-distance connections across LATA boundaries. IXC networks are connected to the LECs through a Point of Presence (POP) which typically is in the form of a tandem switch. A POP is a location where IXC interfaces BOC for exchange access to IXC services. The IXC POP is connected to the LEC access tandem switch via dedicated trunks leased from the LEC. Alternatively, the IXC may collocate network termination equipment in the LEC office, assuming that space is available and that secure physical separation can be established and maintained. IXCs provide inter-lata services.

LEC Domain

IXC Access Types

IXC Domain

Direct Access
Switch
Tandem Switch

Switched Access
Customers

Central end office Central Tandem office Remote Terminal (RT)

LEC End Office

Switch

Tandem Switch

POP

POP
Tandem Switch

Customer PBX

POP

Switch
LEC Access Tandem

Customers

LEC End Office

POP

Distribution Network

Feeder Network

Access (Local) Network

Regional Network

Long-distance Network

Customer has large enough volume of traffic accessing the POP or requiring egress from it to pay for the direct connect facility, bypassing the LEC switching network.

Customer traffic to/from POP doesnt justify direct connect. The IXC purchase access/egress facilities from the LEC which uses its switched network to deliver/receive that traffic.

Basic Architecture of a PSTN

End user access to an IXC via a CAP, bypassing the LEC

IXC domain
ATT POP

Achieving Connectivity

CAP Fiber Ring

Sprint POP

Office Park

Switch

MCI POP

Full Mesh
CAP

Shared Medium

Role of Switching
Connectivity, network resource sharing, customer coordination Dedicated Line Time Shared Synchronous TDM

Sharing Transmission Bandwidth

Time Shared Packet, Burst

Circuit Switching Circuit refers to the capability of transmitting one telephone conversation along one link. To set up a call, a set of circuits has to be connected, Joining the two telephone sets. By modifying the connections, the operators can switch the circuits. Circuit switching occurs at the beginning of a new telephone call. Operators were later replaced by mechanical switches and, eventually, by electronic switches. An electronic interface in the switch converts the analog signal traveling on the link from the telephone set to the switch into a digital signal, called a bit stream. The same interface converts the digital signal that travels between the switches into an analog signal before sending it from the switch to the telephone. The switches use a dedicated data communication network Common channel signaling (CCS) to exchange control information among themselves. Thus CCS separates the functions of call control from the transfer of voice.

Circuit Switching Continue

In current telephone networks, the bit streams in the trunks (lines connecting switches) and access links (lines connecting subscriber telephones the switch) are organized in the digital signal (DS) hierarchy. The DS-1 signal carries 24 DS-0 channels, but its rate is more than 24 times 64 kb/s. The additional bits are used to accommodate DS-0 channels with rates that deviate from the nominal 64 because the signals are generated using clocks that are not perfectly synchronized. Since the 1980s the transmission links of the telephone network have been changing to the SONET or Synchronous Optical Network, standard. In circuit switching, the route and bandwidth allocated to the stream remain constant over the lifetime of the stream.

Rate in Mb/s

Circuit Switching Continue

The capacity of each channel is divided into a number of fixed-rate logical channels, called circuits. The division is usually accomplished by TDM. Circuit switching involves three phases: (1) The source makes a connection or call request to the network, the network assigns a route and one idle circuit from each link along the route, and the call is then said to be admitted (if the network is unable to make this assignment, the call is rejected). This phase is called connection setup. (2) Data transfer now occurs-the duration of the transfer is called the call holding time. (3) When the transfer is complete, the route and the circuits are deallocated. That phase is called connection teardown.

Meium T-1 paired Cable T-1C paired cable T-2 paired cable T-3 coax, radio, fiber Coax, waveguide, radio, fiber

Signal DS-1 DS-1C DS-2 DS-3 DS-4

No. of Voice Circuits 24 48 96 672 4032

North America 1.5 3.1 6.3 45.0 274.0

Europe 2.0

8.4 32.0

Digital Signal Hierarchy

Note that the bit rate of a DS-1 signal is greater than 24 times the rate of voice signal (64 Kb/s) because of the additional framing bit required.

Time Division Multiplexing

Channel 1

Synchronous Transfer Mode

Frame 1

Frame 2

Channel 2

1 2

...

...

N
PBX

...
Channel N

Circuits / Time Slots

TDM is ideal for constant bit rate traffic. The capacity of the outgoing channel is divided into N logical channels. Time on the outgoing channel is divided into fixed-length intervals called frames. Frames are delimited by a special bit sequence called a framing pattern. Time in each frame is further subdivided into N fixed-length intervals called slots/circuits. Each frame consists of a sequence of slots: slot 1, slot 2,.., slot N. (A slot is usually 1 bit or 1 byte wide). A logical channel occupies every Nth slot. There are thus N logical channels. The first logical channel occupies slots 1, N + 1, 2N + 1,..; the second occupies slots 2, N+2, 2N+2,...; and so on.

STM Multiplexer Router 3 2 1 3 2 1

Workstation
STM Multiplexing is also known as Time Division Multiplexing (TDM)

TDM Continues
The T1 Frame (or the OSI term, PDU) consists of 24 8-bits slots. The TDM multiplexer operates as follows: The data bits in each incoming channe1 are read into a separate FIFO (first in, first out) buffer. The multiplexer reads this buffer in sequence for an amount of time equal to the corresponding slot time: buffer 1 is read into slot 1, buffer 2 is read into slot 2, etc. If there are not enough bits in a buffer, the corresponding slot remains partially empty. The bit stream of the outgoing channel is easily demultiplexed: the demultiplexer detects the framing pattern from which it determines the beginning of each frame, and then each slot.
Channel 1

Statistical Multiplexing (SM)

Channel 2

...
Channel N

Most effective in the case of bursty input data. As in TDM, the data bits in each incoming channel are read into separate FIFOs. The multiplexer reads each buffer in turn until the buffer empties. The data read in one turn is called a data packet.

Asynchronous Transfer Mode

A Y PBX B ATM Multiplexer Z Router C Workstation Z Y Z Z Z Y

SM Continues
In TDM each FIFO is read for a fixed amount of time-one slot-and so each incoming channel is allocated a fixed fraction of the outgoing channel capacity, independent of the data rate on that channel. By contrast, in SM, the capacity allocated to each incoming channel varies with time, depending on the instantaneous data rate: the higher the rate, the larger the capacity allocated to it at that time. The size of packets read from each FIFO can vary across channels and over time within each channel. The demultiplexer cannot sort the packets belonging to different channels merely from their positions within a frame.

SM Continues
Additional bits, which delimit each packet and identify the corresponding incoming channel or source, must be added to each packet. The resulting overhead is significantly larger than under TDM. Multiplexer and demultiplexer implementations are more difficult; Multiplexer must now add the packet delimiter and channel or source identifier. Demultiplexer must locate and decode those bit patterns. These increases in complexity and overhead must be balanced against high utilization in the face of bursty data to determine whether SM or TDM is more efficient.

DATA COMMUNICATIONS

Data
Binary Codes Between machines, information is exchanged by binary digits (bits). Two sets are in common use today: ASCII: the American Standard Code for Information Interchange employs a sequence of seven bits. Since each bit may be 0 or 1, ASCII contains 128 unique patterns. EBCDIC: the Extended Binary Coded Decimal Interchange Code employs a sequence of eight bits. It contains 256 unique patterns. There are two basic methods of data transmission Asynchronous and Synchronous. Asynchronous (Character Framed) Transmission; Characters are generated and transmitted singly, one after the other. In some terminals, the characters are collected until a complete line of text is created, or the return key is pressed, causing the line to be sent as a burst of continuous characters.

Data Continues
Whether sent one-by-one as they are generated, or sent line-by-line as each line is completed, each character is framed by a start bit (0) and a stop bit (1) Synchronous (Message Framed ) Transmission: Such transmission is message framed and overcome the inefficiencies of asynchronous, start-stop transmission for high speed data transmission. Rather than surrounding each character with start and stop bits, a relatively large set data is framed, or blocked with one or more synchronization bits or bit patterns used to synchronize the receiving terminal on the rate of transmission of the data. The start sequence is called the header it contains synchronizing, address, and control information. The stop sequence is called the trailer it contains error checking and terminating information. The entire data entity is called a Frame

Character In asynchronous transmission, each character is framed by one start bit and one or two stop bits. Stop Bit (1) Start Bit (0)

Asynchronous Transmission Format

Error Control/Detection
Message Datastream that includes redundant bits and the result of the senders calculations Message

Sender

Receiver

Framed characters sent as they are created -- a data stream typical of keyboard input to a terminal or communications controller.

Sender adds redundant bits and performs calculations to assist the receiver in error detection

Receiver checks redundant bits and repeats calculations looking for agreement with senders results

Framed characters that are concatenated and sent when a string is completed -- a datastream typical of a terminal sending keyboard input lineline-byby-line to a communications controller
Data Block

Because each character is assigned a unique code, it is extremely important to be sent without error. For instance, the ASCII code for p which is the is 1110000. 111000 1110000 An error in bit # 1produces 1110001 code for q. Error detection is a cooperative activity between the sender and the receiver in which a sender adds information to the character or frame to assist the receiver in determining whether an error has occurred in transmission or reception.

Trailer
Frame

Header
Character

Synchronous Transmission Format

Characters are assembled into a datablock that is framed by a header and a trailer to produce a frame. The frame is sent when a command is received from the controlling unit in the communication system.

Error Correction
Sender performs calculation... Receiver performs same calculation...

MK Gn+1

MK = integer + Fn Gn+1

= integer + Fn

If Fn = Fn transmission is without error If Fn Fn transmission is without error

MK

Sender adds Frame Check Sequence (Fn) to frame Gn+1 Generating Function

Fn MK

Receiver re-calculates Fn Gn+1 Generating Function

MK

Once detected,an error must be corrected. Two basic approaches to error correction: 1. Automatic-Repeat-Request (ARQ): Requires the transmitter to re-send the portions of the exchange in which errors have been detected. ARQ techniques include: Stop-and-Wait: The sender sends a frame and waits for acknowledgement from the receiver. This technique is slow. Go-back-n: 2. Forward Error Correction (FEC): FEC techniques employ special codes that allow the receiver to detect and correct a limited number of errors without referring to the transmitter. This convenience is bought at the expense of adding more bits (more overhead)

Cyclic Redundancy Check

MODEM Data Terminal Equipment

DTE

EIA232

DCE

Analog (Voice Grade) Line

Data Communication
Data Circuit Terminating Equipment (DCE): is the equipment that interfaces the DTE to the network; maps the incoming bits into signals appropriate for the channel, and at the receiving end, maps the signals back to bits. DCEs includes modems, digital service units (DSUs), and channel service units (CSUs).

Digital Signals

Data Circuit Terminating Equipment

DTE

EIA232

DSU/CSU

Digital Line

If the transmission channel is an analog line (voice-grade), the DCE is called a modem. When sending, DCE convert the digital signal received by the DTE to analog signals to match the bandwidth of the channel. If the connections are digital connections, the DCE consists of two parts: DSU- receives unipolar digital signals from the DTE and converts them to bipolar signals. CSU: provides loopback (for testing), limited diagnostic capabilities. When sending, it converts bipolar signals to AMI.

The data equivalent of Customer Premise Equipment (CPE) in the voice world, Data Terminal Equipment (DTE) comprises the computer transmit and receive equipment; are digital devices that send or receive data messages. Internally, their signals are simple, unipolar pulses; externally, they may use one the more sophisticated digital signaling schemes.

Protocols
Data Communication Continues
EIA232 interface A DET is connected to a DCE by a cable that conforms to EIA232 standard. EIA232 describes a multi-wire cable that terminates in 25-pin connectors. The cable supports asynchronous or synchronous operation at speed up to 19.2 kb/s. At 19.2 kb/s, the cable length is limited to 50 feet. The EIA232 circuits linking DTE and DCE carry signals that initiate, maintain, and terminate communication between the two. Higher Speed Interconnections EIA449: It permits operation up to 2 Mb/s at distances up to 4000 feet. Enterprise Systems Connection (ESCON): an optical fiber connection operating up to 40 kilometers at 17 Mb/s. Fiber Channel Standard (FCS): Operates up to 10 kilometers at speeds up to 800 Mb/s. FCS includes error control and switching.

Data Link Control (DLC) Protocol A set of rules that governs the exchange of messages over a data link. DLC protocols are divided into two classes: Asynchronous Operation: Start-Stop DLC protocol synchronous Operation: Bit-oriented DLC protocol (e.g., SDLC): Introduced in 1972, SDLC was modified and standardized by ITU-T and ISO as:

HDLC (High Level Data Link Control Protocol) LAP-B (Link Access-Procedure Balanced), for X.25 Standard LAP-D ((Link Access-Procedure Channel), for ISDN-D Channel LAP-F ((Link Access-Procedure Frame Relay), a version of LAP-D used in Frame Relay applications.

Different in the detailed meaning of specific control field bits, all of these protocols share a common structure. In the order that they are transmitted, they consist of the following fields: Flag, Flag Address, Address Control, Text Frame Check Sequence, Sequence and Flag. Flag Contro Text,

SDLC Frame Format

Start Bit

CHARACTER ASCII a

Stop Bit

Start Bit

CHARACTER ASCII b

Stop Bit

01111110

01111110

NR

Receive Sequence Number Number (in sequence 000 through 110) of frame

SDLC FRAME
Header Trailer

Line Idle 0 1 0 0 0 0 1 1 1 State

Timing Mark

Control

Line Idle State

Line 0 0 1 0 0 0 1 1 1 Idle State

Timing Mark

F L Address A G 8 bits 24

TEXT usually 1024 bits

(not Supervisory Frames) Nx8

F C S
16

F L A G 8

expected. 111 acknowledges sequence of seven frames. NS Send Sequence Number Number (in sequence 000 through 110) of this

Time between characters

Information Format

Transmission Format for Start - stop (Asynchronous) Signaling. In idle state, the line is maintained at the 1 level. The start bit (0) reduces the level to zero signaling the commencement of activity.

frame. Mode 00 = Ready to Receive 10 = Not ready to Receive

0 1 0

NS

F NR P NR

Supervisory Format
Mode

01 = Reject

NO TEXT P = 0 = not polled 1 = poll

F = 0 = more frames to come. Information transfer is not complete. 1 = last Frame

Packet Switching
The data stream originating at the source is divided into packets of fixed or variable size. The time interval between consecutive packets may vary, depending on the burstiness of the stream.

PACKET SWITCHING

As the bits in a packet arrive at a switch or router; they are read into a buffer when the entire packet is stored, the switch routes the packet over one of its outgoing links. The packet remains queued in its buffer until the outgoing link becomes idle. This store-and-forward technique thereby introduces a random queuing delay at each link; The delay depends on the other traffic sharing the same link. Packets from different sources sharing the same link are statistically multiplexed.

Packet Switching Continues

The routing decision
Connectionless (datagra Connectionless (datagram)

Connection-Oriented vs Connectionless Transport

Connection Oriented Circuits and Virtual Circuits Guaranteed Resource Connection State Delay Yes Constant Guaranteed Busy Maintained Shared Resource Yes Variable Shared Share Pain Maintained No Variable Shared Share Pain Could change

Connection Oriented (virtual circuit)

Connectionless

In datagram packet networks, each packet within a stream is independently routed. A routing table stored in the router (switch) specifies the outgoing link for each destination. The table may be static, or it may be periodically updated. Each packet must contain bits denoting the address of the source and destination. In virtual circuit packet networks, a fixed route is selected before any data is transmitted in a call setup phase similar to circuit-switched networks. However; there is no notion of a fixed-rate circuit or logical channel. All packets belonging to the same data stream follow this fixed route, called a virtual circuit. Packets must now contain a virtual circuit identifier; this bit string is usually shorter than the source and destination address identifiers needed for datagrams. However; the call setup phase takes time and creates a delay not present in datagram packet networks.

Bandwidth Overload Packet Sequence

Connection Oriented Packet Transport

Connectionless Transport

Connection Request Resource Check Route Selection Destination Acceptance Connection begins

Lower Level Protocol (IP) Send and Pray Upper Level Protocol Guaranteed delivery

Circuit Switching Techniques

Message Switching

Packet Switching

Frame Relay (Switching)

Cell Relay (Switching)

Fast Relay

Relay

Direct Connection

Store & forward

Hold & forward

Hold & forward

Hold & forward

Frame Relay
(Variable size PDUs--frames)

Cell Relay
(Fixed size PDUs-cells)

Copper, wireless

Copper, wireless

Copper, wireless, optical

Copper, wireless, optical

Copper, wireless, optical

Media

PVC
Size of No such thing Variable, large to small Variable, large to small Variable, large to small PDU

SVC (Q.931)

ATM Based (For B-ISDN)

802.6 Based (For SMDS)

(LAPD)
Fixed, very small

PVC
Delay Very Fast Slow Fast Faster Very Fast

SVC (Q.2931)

Switching Technologies

Types of relay systems

Typical X.25 Topology

X.75 (NNI) X.25 X.25

X.25 Continue
X.25 encompasses the lower three layers of the OSI model

User

User

= Packet switches

X.25X.25-3 layer (network layer) Packets are created at the network layer that Establishes, manage, and teardown the connections between the user and the network. X.25X.25-2 layer (data link layer) The packet is encapsulated within the Link Access Procedure, Balanced (LAPB) protocol as the information field. The LAPB protocol is a subset of HDLC (High Level Data Link Control). X.25X.25-1 layer (physical layer) The physical layer is the physical interface between the DTE and the DCE.

X.25 is not a packet switching specification. Its a packet network network interface specification. X.25 says nothing about operations within within the network.

It Provides for an interface between an end-user device (DTE) and a network (DCE). Its formal title is Interface between DTE and DCE for terminals operating in the packet node on public data networks In X.25, the DCE is the agent for the packet network to the DTE.

X.25 Continue
X.25 uses logical channel numbers (LCNs) to identify the DTE connections to the network. An LCN is really nothing more than a virtual circuit identifier (VCI). Octets #1 and Octet #2 of the packet header provide a 12-bit identifier. If all-zeros possibility is excluded, as many as 4095 logical channels (i.e., user sessions) can be assigned to a physical channel. The LCN serves as an identifier (a label) for each user's packets that are transmitted through the physical circuit to and from the network. Typically, the virtual circuit is identified with two different LCNs-one for the user at the local side of the network and one for the user at the remote side of the network. X.25 provides two mechanisms to establish and maintain communications between the user devices and the network (and ATM has borrowed these concepts): Permanent Virtual Circuit (PVC) and Switched Virtual Circuit (SVC).

X.25 Continue
PVCs may support large users. All packets travel the same path between two computers; which path is established by routing instructions programmed in the involved nodes. The circuits involved in the route are defined on a permanent basis, until such time as they are permanently redefined, perhaps as the service Alternatively, the network may select the most available and appropriate path on a callby-call basis using Switched Virtual Circuits (SVCs); Again, all packets in a given session travel the same path. SVCs demand a greater level of network intelligence that adds to total network cost; this translates into higher cost to the end-user organization. The establishment of a SVC also involves some level of delay since the network nodes must examine multiple paths in order to make a proper selection.

USERS INFORMATION
USER-NETWORK INTERFACE X.25 User Stack
Users Data I.e. message data and/or headers from upper layers

PACKET NETWORK
Users Data Segment Users Data Segment Users Data Segment Users Data Segment

Transport
Packet Header Header

HDLC FRAME Packet

Trailer

Packet X.25-3 LAPB X.25-2 X.21 X.25-1

LAPB Header

Packet Data LAPB Trailer Data

Network
LAPB X.21 Data Link Physical

F L A G

A d d r e s s

C o n Packet t r Headers o l

Users Data Segment 1024 bits

FCS

F L A G

DTE

DCE

Logical Grp # 1 0 D Q
Logical Channel Number

0 P(S)

P(R)

X.25 Packet and Frame Format

RS-232-C (1969)

2.4 38 Kbps

COMPUTER NETWORKS

01101011_11011010_

The RS-232-C standard for the serial line specifies the transfer of one 8-bit character at a time, separated by time intervals. The speed and distance of the serial line are limited.

The Synchronous Data Link Control and related standards transmit long packets of bits. The header (H) contains the preamble that starts the receiver clock, which is kept in phase by the self - s ynchronizing encoding of the bits. The receiver uses the cyclic redundancy check (CRC) bits to verify that the packets is correctly received.

StoreStore-andand-forward transmissions proceed by sending the packet successively along links from the source to the destination. The packet header specifies the source and destination addresses (A and E, for example) of the packet. When it receives a packet, a computer checks a routing table to find out on which link it should next send the packet.

A 4 or 16 Mbps

C D

E C D E

Ethernet. In this network, computers are attached to a common coaxial cable. The computers read every transmitted packet and discard those not addressed to them.

Token ring. The computers share a ring. Access is regulated by a token - p assing protocol.

155-622 Mbps A 100 Mbps E D B A B

Fiber Distributed Data Interface (FDDI). A tokentoken-passing protocol is used to share the ring. The computers time their holding of the token. This network guarantees that every computer gets to transmit within an agreedagreed-on time.

Asynchronous Transfer Mode (ATM) network. The network transports information in 5353-byte cells. Total throughput of this network is much larger than that of FDDI or of a 100100-Mbps Ethernet.

CPU

Display

CACHE

VRAM

CPU

RAM User System

LAYERING APPROACH

DISK

RAM

Keyboard, mouse, etc.

NIC NIC

Computer Message Transfers The left panel gives a simple architecture of a host computer and its connection to the network. The right panel shows the four copies that may be involved across the CPU bus to run an application, reducing the host throughput.

OSI Hierarchy
7 6 5 4 3 2 1 Application Presentation Session Transport Network Link Physical

OSI Hierarchy
7 6 5 4 3 2 1 Application Presentation Session Transport Network Link Physical

Physical
SONET, T1, T3

Transport
Error and congestion control TCP, UDP

Link
Ethernet, FDDI Circuit, ATM, FR switches

Session, Presentation, Application

Data, voice encodings Authentication web/http, ftp, telnet

Network
Routing, Call control IP internetworking

Data Transfer Over Frame-based Networks

File

Data Transfer Over Cell-based Networks

File

TCP

TCP

IP IP

Adaptation Frame (Ethernet, FR, PPP)

ATM Cells

Internet Protocol Architecture

Ping Ping FTP FTP TELNET TELNET HTTP HTTP DNS DNS RTP RTP SNMP SNMP

Why a Synchronous Network

Visibility of each byte at the line rate Simplification of the multiplexing and switching process
OSPF OSPF

SMTP SMTP ICMP TCP TCP

BGP BGP

RIP RIP UDP UDP

Simple access to overhead bytes Asynchronous

IP LANs LANs 10/100BaseT 10/100BaseT ATM ATM FR FR PPP PPP Circuit-Switched Circuit-SwitchedB/W: B/W: POTS, POTS,SDS, SDS,ISDN, ISDN,... ... CDPD CDPD Wireless Wireless

Stuffing Bits
OH OH

Synchronous

Dedicated DedicatedB/W: B/W: DSx, DSx,SONET, SONET,... ...

Overhead functions framing, monitoring, fault location, protection switching, management communications.