SkyEdge Access Schemes - 0408
SkyEdge Access Schemes - 0408
SkyEdge Access Schemes - 0408
April 2008
This document contains information proprietary to Gilat Satellite Networks Ltd. and may not be
reproduced in whole or in part without the express written consent of Gilat Satellite Networks Ltd. The
disclosure by Gilat Satellite Networks Ltd. of information contained herein does not constitute any
license or authorization to use or disclose the information, ideas or concepts presented. The contents of
this document are subject to change without prior notice.
SkyEdge Inbound Access Schemes
Contents
Figures
Figure 1: SkyEdge VSAT Random Access (RA) mode .....................................................4
Figure 2: RA frequency plan.............................................................................................4
Figure 3: CRA access scheme .........................................................................................5
Figure 4: CRA Grades of Service .....................................................................................6
Figure 5: GA/RRA Access Scheme ..................................................................................7
Figure 6: SkyEdge Dedicated Access (DA) mode.............................................................9
Figure 7: Normal vs. Rain Fade mode ............................................................................23
Figure 8: Star Topology..................................................................................................24
Figure 9: Mesh Topology................................................................................................25
Tables
Table 1: CRA Grades of Service ......................................................................................6
This document provides information about SkyEdge Version 5.1 access schemes and
indicates which access methods are appropriate for various types of applications.
Audience
NOTE
Screen captures in this document were obtained using SkyEdge Version 5.1.
These screens vary in other versions but the differences are not significant
for the purposes of this document.
In most access schemes, inbound resources are managed and controlled by the
SkyEdge hub. Some access schemes, such as RA (Random Access), do not require
inbound resource management. Others, such as GA (Guaranteed Access), do require
the assignment of inbound resources by the hub.
Most SkyEdge VSATs have an access layer (RSP/LAPU), a data application layer
(DRPP) and a telephony application layer (VRPP). The access layer provides the
following functions:
Transmits application traffic over the channel
Ensures data reliability
Initiates, modifies, and releases sessions
Can reserve bandwidth based on a combination of the following criteria:
− Volume vs. rate-based reservation
− Allocation granularity
− Control channel requirements (capacity requests)
A single SkyEdge HSP can support several satellite networks. Each network has its
own inbound RF resources except for physical time-slot duration which is shared by
the separate inbound bands.
Each VSAT is allocated to one inbound band. Each inbound band includes channels
that VSATs use for sending signaling to the hub and for transmitting data to the hub
or to other VSATs when Mesh is used.
Managed Private Networks (MPNs) enable several satellite networks to share the
same RF resources. Unlike inbound bands, MPNs use the same inbound space
segment for all of the VSATs in the network. Each MPN enforces logical restrictions
regarding bandwidth allocation to VSATs. This enables different bandwidth
guarantees and limitations to be applied to each MPN.
QoS features that are not related to any access-scheme can be implemented at the
VSAT level. These QoS features can prioritize the output sequence of buffered
traffic at the VSAT based on application importance or priority regardless of the
amount of bandwidth that is allocated to the VSAT.
SkyEdge uses a combination of satellite access methods for inbound and outbound
transmission. This suite of access modes enables the network to be optimized for the
customer’s needs and applications. VSATs use the inbound (return) channels for
transmission to the hub (Star topology) and towards each other (Mesh topology).
If a collision occurs, data packets are retransmitted (after receiving a no-ack within a
specified time-interval) at a randomly chosen frequency in a subsequent time slot to
minimize the chance of a second collision. This mode is well-suited for interactive
traffic characterized by short messages and light channel loads.
Multiple VSATs share the same satellite channel inbound bandwidth on a Time
Division Multiple Access (TDMA) basis using a modified Slotted ALOHA access
scheme. Not only does an inbound VSAT carrier burst in different time slots, it also
bursts in random frequency slots. This two dimensional (time and frequency) access
approach provides natural balancing of VSAT inbound traffic loading.
11 31
14
21
21 14
14
25 14
At any time, over 80% of the traffic on a satellite network is generated by a relatively
small number of “busy” VSATs. These busy VSATs generate the majority of
collisions. To deal with this phenomenon Gilat developed and patented the CRA.
Busy VSATs typically remain busy for 100 milliseconds to three minutes. CRA
identifies the busiest VSATs and ensures that their transmissions do not collide with
each other by distributing their traffic across the inbound bandwidth. This increases
slot utilization for the RA protocol, raises the activation threshold for the DA
protocol, reduces the total number of collisions, and increases throughput.
Using real-time heuristic algorithms, the HSP identifies the busiest VSATs and
assigns each one to a separate channel or part of a channel in the RA channel range.
Each busy VSAT can transmit only in its specific slot segment. After a period of
time, the HSP again determines which VSATs are busiest and repeats the process.
CRA does not guarantee a collision-free environment because non-busy VSATs can
collide with busy and other non-busy VSATs. However, ensuing that the busiest
VSATs do not collide with each other reduces collisions by about 80%.
If, for example, 9 VSATs are identified as busy, Figure 4 shows how CRA divides
these VSATs into 3 grades of service. Table 1 describes these grades of service.
The CRA access scheme is optimized for the following customer applications:
Networks with large numbers of VSATs
Networks that experience congestion but cannot afford reservation channels
IP traffic that can accommodate a medium-to-high amount of jitter
Applications that require short response times
Applications that require simple deployment procedures
Applications that require the network to automatically adapt to user needs
A VSAT uses the RRA protocol for transmitting initial control messages and for
requesting GA time slots based on the number of inbound transmissions waiting in
its queue. Data is transmitted over the allocated time-slots using the GA protocol.
The hub receives VSAT requests and allocates one or more contiguous time slots in
an available GA channel. The allocation contains the frequency, the first available
starting time slot, and the number of time slots. Because no two VSATs ever receive
the same allocation, collisions and retransmissions do not occur when using GA.
The RRA protocol VSAT request-hub allocation exchange causes a relatively long
650 millisecond delay before the first data packet reaches the hub. However, for long
transactions and streaming applications, the effect of this delay is minimized by
piggy-backing subsequent allocation requests on the GA protocol data packets.
The RRA protocol directs allocation requests and manages the GA protocol. After a
VSAT receives a GA allocation, the VSAT transmits on a single frequency over a
continuous span of time slots. Collisions and retransmissions do not occur when
using GA because no two VSATs ever receive the same allocation.
The GA/RRA access scheme is optimized for the following types of applications:
Applications that require high channel utilization
IP traffic that can accommodate a medium-to-high amount of jitter
Applications that require network-managed bandwidth
Applications that require differentiated access services
Consumer satellite broadband internet access
NOTE
RA and DA protocols cannot exist on the same carrier.
VSATs that use both protocols must use at least two carriers.
VSATs in DA mode are assigned a specific frequency and time slot (TS) to avoid
contention with VSATs in RA mode. VSATs use DA when continuous volume or
periodic assignment is required between a remote site and the hub (for example, to
upload an FTP file).
11 34 31 17
14 17
21 17 34
21 14 17 34
25 17 34
17 34
25 14 17 34
17 34
DDA measures the rate of all traffic going through the VSAT and automatically
issues a capacity request that provides enough bandwidth for that traffic.
The reservation model measures the user demand (bandwidth needs) and generates
capacity requests accordingly. As in GA, the hub uses advanced algorithms to
distribute network bandwidth to all requesting VSATs based on amount demanded,
VSAT priority and SLA requirements.
− Uses VSAT priority classification (e.g., platinum, gold, silver and bronze)
The DDA access scheme is optimized for the following types of applications:
RA is suitable for networks with a large number of VSATs whose average bandwidth
demand is low and whose applications need a very short response time (about 650 ms
or approximately the time required for one satellite round-trip). Because RA is based
on random access, it is economical in situations where VSAT average throughput is
low so that there is a high probability that bursts do not collide.
3.1.4. Limitations
RA is economic only at high contention ratios.
Typically, RA provides adequate response time if not more than 33% of the
available bandwidth is utilized. Therefore, IB bandwidth should be about three
times the overall expected IB demand.
RA is not suitable for multimedia applications because guarantees are not
provided for absence of delay or availability of a jitter-free IB channel.
RA is not suitable for guaranteed applications where, for some reason, a specific
response time must be provided 100% of the time. RA usually (but not always)
provides good response time.
RA can provide satellite access to applications where the number of remote terminals
is very high if the bandwidth needed by each terminal is very low. RA economically
provides adequate application bandwidth and response time without reserving any
bandwidth for each remote terminal.
3.2. CRA
CRA is appropriate for networks with many VSATs some of which have throughput
requirements that exceed the level suitable for RA.
3.2.3. Limitations
Typically, CRA can provide adequate response time if not more than 50% of the
available bandwidth is utilized. Therefore, IB bandwidth should be about two
times the overall expected IB demand.
CRA is not suitable for multimedia applications because guarantees are not
provided for absence of delay or availability of a jitter-free IB channel.
CRA is not suitable for guaranteed applications where, for some reason, a
specific response time must be provided 100% of the time. CRA usually (but not
always) provides good response time.
CRA can provide satellite internet access for large numbers of customers who share
the same bandwidth pool. Although internet traffic patterns show little stability over
time, CRA minimizes satellite resource usage by reducing collisions between current
high-traffic users.
3.3. RA + DA
This access scheme combination is suitable for networks with many VSATs each of
which has low traffic requirements (as in RA). However, at specific times some
VSATs need a heavier dedicated channel for specific applications. These
applications are detected and trigger a switch to DA for the duration required.
The switch to DA is triggered based on application profiles
DA triggers include detection of TCP and UDP ports or VoIP sessions
The network sizing process manages two bandwidth segments: RA for normal traffic
and DA for specific applications.
RA access minimizes bandwidth usage when VSAT demand is low but still provides
very good average response time. DA access is used when RA access is not adequate
for VSAT bandwidth requirements.
3.3.3. Limitations
For RA limitations, see Section 3.1.4.
For DA limitations, see Section 2.4.1.
Most ATM network traffic runs over RA because sustained traffic rates are very low.
However, log files are uploaded once a day by running file transfer applications over
DA channels. DA minimizes the upload time because the remote terminal does not
have to deal with collisions.
3.4. RRA + GA
3.4.3. Limitations
Long response times.
GA is not suitable for multimedia applications because guarantees are not
provided for absence of delay or availability of a jitter-free IB channel.
GA is not suitable for guaranteed applications where, for some reason, a specific
response time must be provided 100% of the time. GA usually (but not always)
provides good response time.
RRA + GA can be used for satellite internet access where many customers share the
same bandwidth pool. The volume-based reservation maximizes space-segment
utilization and GA provides differentiated SLA.
BoD is useful when a medium/small network of VSATs have both strict response-
time and SLA/QoS requirements.
BoD uses RA resources for VSATs that currently have low traffic and DA resources
for VSATs that currently have high traffic. Unlike other access schemes, both RA
and DA are dynamically allocated based on the VSAT’s actual bit-rate needs.
3.5.3. Limitations
Not suitable for large networks (more than a few hundreds of VSATs)
BoD is used for satellite internet access or enterprise networks with a medium
number of sites that share the same bandwidth pool. The rate-based reservation
maximizes space-segment utilization and provides differentiated SLA that supports
jitter-sensitive applications such as VoIP and mesh connectivity in the same
bandwidth pool.
The main advantages of the BoD access scheme are optimized utilization of the
reserved DA resource, fair and efficient BW distribution, and the fact that QoS is
used at both the VSAT (CIR/MIR) level and MPN level. In addition, BoD can
efficiently serve both star and mesh links simultaneously.
BoD allows defining a wide range of possibilities for each VSAT’s service level. The
following two examples illustrate practical scenarios using BoD:
VSATs that are sensitive to response time can be guaranteed a constant bit rate
(CBR) by constantly allocating a small number of DA slots. These VSATs use
RA channels for network logon and then permanently move to DA where their
allocation level varies between the CBR and VSAT MIR levels.
BoD enables VSATs to be assigned low constant bit rates (CBR) and enjoy constant
and guaranteed bandwidth with fast response time.
Using the CBR option in BoD, a network operator can provide the SCPC market with
a cost-effective solution based on SkyEdge. Although the utilization of the SCPC
space segment is often far from optimal, dynamic reserved allocation of the space-
segment allows significant reduction in the expenses of SCPC operators, with
minimal and controllable compromise on VSAT performance.
Multiple Inbound Rates – Different bit rates can be used for different types of
channels in the same inbound band. Each of these groups can have a different rate:
RA/RRA/GA
DA
Star VDA (SVDA)
Mesh VDA (MVDA)
Rates of X, 2X and 4X can be assigned to the different groups, subject to the link-
budget. Different inbound bands can use different unrelated rates.
All channels in the same inbound band use the same FEC.
If the flexibility provided by multiple inbound rates is not sufficient, multiple IBBs
can be used. Multiple inbound bands enable groups of VSATs to use different RF
resources, different channel rates, and different access-schemes. This enables the
network to be optimized differently for separate groups of VSATs.
However, all inbound bands running on the same HSP must use the same time-slot
duration.
VoIP support provides simple, IP-based solutions that permit the same bandwidth to
be shared between VoIP and other network traffic. However, the VoIP solution
requires special voice-terminal equipment (VoIP gateways / ATA) to carry voice
over the IP network.
4.4. Mesh IP
Mesh link allocations are managed asymmetrically based on the amount of data each
party needs to transmit. For real-time applications (e.g., VoIP), capacity is allocated
symmetrically for both sides.
Mesh Limitations:
HTTP acceleration, Multicast, and Virtual Private Networks are not supported.
Mesh traffic is untagged. Therefore, the VSAT cannot be assigned to a VLAN.
Providing cellular telephony in remote and rural areas can be very costly because
terrestrial infrastructure is limited or non-existent. Gilat’s SkyAbis solution provides
a backhaul link between each remotely deployed BTS (Base Transceiver Station) and
a central BSC (Base Station Controller). Usually a single BSC handles multiple BTS
units, only some of which are satellite-accessed. However, it is possible to backhaul
an entire BSC from an MSC (Mobile Switching Center).
When a cellular end user dials a number, the BTS opens a new channel in a TRX and
transmits the digital voice and E1 signaling to the SkyAbis Adapter. The SkyAbis
Adapter compresses the digital voice signal, encodes it into IP packets, and sends the
voice and signaling packets over a UDP/IP LAN interface to the VSAT. The SkyAbis
Adapter at the hub de-multiplexes and converts the packets back into E1 format and
forwards the digital voice and signaling to the BSC.
The SkyAbis network is based on Dynamic Dedicated Access (DDA). The VSAT
automatically determines the required traffic rate and reserves the necessary
bandwidth from the hub to ensure adequate satellite resources.
5.1. Synchronization
Satellite inbound transmission uses a slotted Aloha protocol where all the VSATs
and the Hub are synchronized. A special synchronization mechanism ensures that
every transmission occurs on a time slot boundary. Gilat employs a patented
technology for synchronization over a DVB-S channel). Several mechanisms
maintain the VSAT time-slot synchronization. One of these compensates for the
propagation delay variance caused by satellite movement. Another mechanism
automatically adjusts the timing loop of the VSATs, regardless of their geographical
distance from the hub. This timing loop correction simplifies the installation process
by eliminating the requirement to determine and configure geographic location.
DA, Star VDA and Mesh VDA can share the same channel(s) when using the same
channel rate. Each time-slot is dynamically assigned to one of the sub-channel types.
Fading can be caused by rain or antenna movement due to strong winds in the VSAT
area. The satellite network Inbound link budget includes safety factors that mitigate
temporary fading phenomena. These safety factors cause the VSAT to transmit a
stronger signal or receive inbound transmissions at a slower rate so that Eb/N0
remains above the minimum required for the receiver. The result can cause additional
ODU and antenna costs or reduced network availability.
6. Network Topologies
Star topology is the basic and most commonly used topology. All inbound
transmissions terminate at the Hub.
The SkyEdge system supports also the full mesh topology (with dynamic or
dedicated bandwidth assignments) where VSATs communicate directly with each
other (single satellite hop). The VSATs use a dedicated mesh receiver expansion card
(RBRs) enabling it to directly receive inbound transmissions from other VSATs.