ATM

1
 Ref

 Connection oriented network:

SONET/SDH, ATM, MPLS and optical
networks by Harry Perros

2
 ATM: What it is

 Asynchronous Transfer Mode

 A low-layer networking technology based
on fast packet-switching of small fixed size
packets called cells
 ATM provides a single transport mechanism
for integrated services traffic: data, voice,
video, image, graphics...
 All statistically multiplexed at ATM layer

3
 ATM: What it isn’t

 Synchronous Transfer Mode (STM)

 STM relies on pre-assigned “slots” for each
user within a frame, and global timing
information to mark frame boundaries
 Example: T1 transmission (1.544 Mbps)

4
 T1 Transmission
Framing bit
Room for 24 calls, with 8 bits from each

8 bit sample for call i

24 x 8 bits = 192 bits + 1 framing bit = 193 bits

125 microseconds
(8000 cycles/sec)

An example of one frame from T1 digital transmission scheme

193 bits/frame X 8000 frames/sec
5 = 1.544 Mbps
 T1 Transmission (Cont’d)

Frame 1 Frame 2 Frame 3

Call i uses slot i in each frame

8 bits/slot X 1 slot/frame X 8000 frames/sec
= 64000 bits/sec = 64 kbps
Calls can be allocated k slots per frame to give
allocated bandwidth that is k x 64 kbps
Idle slots are wasted
6
 Synchronous Transfer Mode

 STM relies on positional association: slots

are identified by their relative position from
the start of the frame (global timing info)
 Each user knows which slot(s) to use
 All slots are the same size (e.g., 8 bits)
 Bandwidth allocated in multiples of slots
 Efficient for Constant Bit Rate traffic
 Inefficient for Variable Bit Rate traffic
7
 Asynchronous Transfer Mode

 ATM does not use a priori assignment of

slots to users
 Slots are assigned “on demand” on an as
needed basis
 Users can use whichever slots are empty

8
 ATM Transmission

Slotted transmission scheme

All slots are the same size (53 bytes = 1 ATM cell)
Any user can use any empty slot
No notion of specific slots assigned to specific users
Bandwidth allocation in ATM may reserve a certain
percentage of the total slots for a given call, but which
slots you get is determined at time of transmission
Can support arbitrary bit rates
9
 Asynchronous Transfer Mode

 No global timing relationship between slots

(i.e., cells) of different users (asynchronous)
 Efficient for Variable Bit Rate traffic
 Implication: the cell in each slot has to be
completely self-identifying (i.e., overhead)

10
 Advantages of ATM

 Better for bursty traffic (i.e., VBR)

 Statistical multiplexing gain
 Better network utilization
 Same mechanism works for all traffic types
 Simple and fast hardware switching

11
 Characteristics of ATM

 Point to point technology

 Connection-oriented: an end-to-end
connection (called a virtual channel) must
be set up using a signalling protocol before
any data cells can be sent on that VC
 “Bandwidth on demand”
 Statistical multiplexing
 Integrated services
12
 ATM Cell

 53 bytes
 5 byte header
 48 byte payload (data)
 Virtual Path Identifier (VPI)
 Virtual Channel Identifier (VCI)
 Simple control fields

13
ATM Cell Format

GFC VPI
VPI VCI
VCI
VCI PT RES CLP
HEC

Payload
(48 bytes)

ATM UNI Cell Specification

14
 Why ATM will win

 ATM is a scalable technology

- scalable in bandwidth
- scalable in distance
- scalable in deployment

15
 ATM: Scalable Bandwidth

 ATM is not tied to any particular bit rate or

physical layer network technology
 ATM is simply the abstract concept of fast
packet switching with small fixed size cells
 Can do low speed ATM (e.g., 1.5 Mbps)
 Can do high speed ATM (e.g., 155 Mbps)
 Primary interest: high speed ATM networks

16
 ATM: Scalable Distance

 ATM can be used for LANs

 ATM can be used for MANs
 ATM can be used for WANs
 Initial market: ATM LANs, enterprise area
networks, LAN backbones
 Future: wide area network backbone, ATM
to the desktop, wireless ATM

17
 ATM: Scalable Deployment

 Emphasis on interoperability, compatibility

 Incremental evolutionary path to ATM
 Ethernet => switched Ethernet => ATM hub
 Start with one switch, N ports, plus NICs
 Add more ports as needed
 Add more switches as needed
 Hierarchical cascading structure

18
 Why ATM will win

 ATM is a scalable technology

- scalable in bandwidth
- scalable in distance
- scalable in deployment
 Global standards (ATM Forum, ITU)
 Billions of dollars invested

19
 ATM Forum

 The ATM Forum is a “standards body” of

vendors of ATM products and services
 Mostly industrial organizations, but some
research institutions are members as well
 Initially half a dozen member companies
 Now well over 700 member companies
 URL: http://www.atmforum.com

20
 ATM Forum (Cont’d)

 ATM Forum has a very aggressive

timeline,much more so than the official
ITU standards body
 Meets every 18 weeks or so
 Short term product focus
 Trying to influence standards for ATM
equipment being deployed or designed

21
 ATM Forum (Cont’d)

 Many proposals come before the Forum,

from companies as well as researchers
 Proposals discussed, evaluated at length
 Membership vote; majority rules
 Fast moving standards process
 Results and decisions disseminate into the
research community very quickly

22
ATM Dictionary

23
 Introduction

 ATM networking is filled with a lingo of its

very own, many of which are acronyms, and
many of which are quite fundamental to an
understanding of what is going on in an
ATM network
 Examples: VCI, VPI, PVC, SVC, AAL,
CBR, VBR, ABR, PCR, SCR, QOS, CDV

24
 Review

 ATM: Asynchronous Transfer Mode

 ATM is a statistical multiplexing technique
for high speed integrated services networks,
based on the fast packet switching of small
fixed size (53 byte) packets called cells
 ATM is a connection-oriented low-layer
networking concept

25
 ATM is Connection-Oriented
 An end-to-end path called a virtual channel
must be set up in advance, using an ATM
signalling (control) protocol, before any
data cells can be sent
 All cells of a virtual channel travel on the
same path
 Cells arrive in the order that they were sent
 Switches must maintain state about the
virtual channels passing through them
26
 Definitions

 Virtual Channel (VC)

- a connection between two communicating ATM
entities (e.g., host-switch, switch-switch)
- set up at time of call arrival
- provides a certain grade of service
(negotiated at time of call arrival)
- cell sequence is preserved

27
 Definitions (Cont’d)

 Virtual Channel Identifier (VCI)

- the label associated with a VC
- 16-bit integer in UNI ATM cell format
- carried in ATM cell header for identification
- note that VCI’s are locally significant only
(i.e., assigned on a per link basis by the ATM
devices at either end of that link)

28
 Definitions (Cont’d)

 Virtual Path (VP)

- a group of virtual channels (VC’s) all travelling
between the same two points in ATM network
- used by the network to simplify provisioning,
resource management, providing different grades
of service, etc.
- “bundles up” traffic heading to same destination

29
 Definitions (Cont’d)

 Virtual Path Identifier (VPI)

- the label associated with a VP
- 8-bit integer in UNI ATM cell format
- carried in ATM cell header for identification
- part of two level addressing scheme in ATM
- note that VPI’s are locally significant only
(i.e., assigned on a per link basis by the ATM
devices at either end of that link)

30
 Definitions (Cont’d)

 VP switch: an ATM switch that deals only

with the VPI’s in cell headers (e.g., a core
switch in middle of a large ATM network)
 VC switch: an ATM switch that deals only
with the VCI’s in cell headers (e.g., access
switch at the edge of an ATM network)
 VP/VC switch: an ATM switch that deals
with both VPI’s and VCI’s in cell switching
31
 Definitions (Cont’d)

 Permanent Virtual Channel (PVC)

- a virtual channel connection (virtual channel)
that is set up on a long term basis (e.g., hours,
days, months, years) by a human operator
- involves statically configuring the
“routing table” in ATM equipment
- done as part of network provisioning in current
ATM network testbeds
- supported by all ATM switch vendors
32
 Definitions (Cont’d)
 Switched Virtual Channel (SVC)
- a virtual channel connection (virtual channel)
that is set up by the ATM signalling protocol
between two communicating ATM entities
- set up on an as needed basis, and torn down
when complete
- short term basis (e.g., seconds, minutes)
- involves dynamically configuring the
“routing table” in ATM equipment
- supported by very few ATM switch vendors
33
 Definitions (Cont’d)
 ATM Adaptation Layer (AAL)
- a protocol for converting between higher layer
protocol data units (PDU’s), such as TCP
packets, IP packets, or JPEG images,
and ATM cells for actual transmission
- defines procedures for segmentation and
reassembly (SAR)
- segmentation: packets to cells (done by sender)
- reassembly: cells to packets (done by receiver)

34
 Definitions (Cont’d)
 Traffic Descriptors
- a numerical specification of the statistical
characteristics of an ATM traffic flow
- used by ATM switches at time of call setup
- specifies Peak Cell Rate (PCR), Sustained Cell
Rate (SCR), Maximum Burst Size (MBS), etc.
- different calls can specify different values for
their traffic descriptor
» e.g., voice: PCR = SCR = 1000 cells/sec
» e.g., data: PCR = 10,000 cells/sec, SCR = 1000

35
 Definitions (Cont’d)
 Quality of Service (QOS)
- a specification of the desired (or acceptable)
grade of service required for a traffic flow
- some traffic is delay-sensitive (e.g., voice)
- some traffic is loss-sensitive (e.g., data)
- some traffic is both (e.g., compressed video)
- some traffic is neither (e.g., LAN emulation)
- QOS requested at time of call setup
- ATM network tries to provide requested QOS

36
 Definitions (Cont’d)
 QOS Parameters
- the parameters that can be specified as part of
the QOS request at time of call setup
- examples: cell loss ratio (CLR), mean cell
delay, maximum tolerable cell delay, cell
delay variation (CDV)
- different calls can specify different values for
their QOS parameters
» e.g., voice: delay < 50 msec, CLR < 0.001
» e.g., data: delay < 1 sec, CLR < 0.000001

37
 Definitions (Cont’d)
 QOS Classes
- generic service classes for ATM traffic
- used to help simplify the management and
support of QOS requirements in ATM networks
- currently there are five proposed classes:
CBR, rt-VBR, nrt-VBR, ABR, and UBR
- class specification determines the order of
service for cells of different VCI’s

38
 Definitions (Cont’d)
 Constant Bit Rate (CBR)
- simplest type of traffic: constant bit rate
(e.g., voice traffic, T1 circuit emulation)
- this is the highest priority class because of the
delay-sensitive (i.e., time-dependent) nature of
the traffic carried
- suitable for periodic (isochronous) traffic
- need to specify only PCR (which equals SCR)
- supported by some ATM switch vendors

39
 Definitions (Cont’d)
 Variable Bit Rate (VBR)
- more complicated type of traffic: the bit rate
varies with time (e.g., compressed video)
- peak bit rate (i.e., short term) may be much
higher than the mean bit rate (i.e., long term)
- must specify PCR, SCR, and burstiness
- next highest priority class(es)
- versions: real-time (rt) and non-real-time (nrt)
- supported by some ATM switch vendors

40
 Definitions (Cont’d)
 Available Bit Rate (ABR)
- unpredictable type of traffic: traffic is willing to
use as much or as little bandwidth as is
available (e.g., Internet traffic, LAN emulation)
- usually variable bit rate, delay-insensitive
- referred to as “elastic traffic” (e.g., ftp)
- next lowest priority class
- supported by few ATM switch vendors
- still under discussion by ATM Forum

41
 Definitions (Cont’d)

 Unspecified Bit Rate (UBR)

- the “bottom feeder” in the ATM food chain
- Minimum Cell Rate (MCR) is zero; guaranteed
nothing, but may get more than this if there
happens to be extra bandwidth available
- lowest priority class
- much discussion at ATM Forum
- may become Guaranteed Frame Rate (GFR)?

42
 Definitions (Cont’d)
 Call Admission Control (CAC)
- a control function in ATM switches that makes
the decision on whether or not to accept a
newly incoming call
- considers traffic descriptor (TD) and quality of
service (QOS) parameters, as well as impact on
the QOS of existing calls in the network
- can be statistical or deterministic
- still an active research topic

43
 Definitions (Cont’d)

 Usage Parameter Control (UPC)

- a control function performed in switches to
“police” ATM traffic flows
- monitors traffic on a VCI basis, to make sure
that it conforms to the declared traffic
descriptor (TD)
- in case of violation, can tag cells with a
violation tag (CLP = 1), discard cells, or abort
the connection
44
 Definitions (Cont’d)

 Cell Loss Priority (CLP)

- a single bit in ATM cell headers for denoting
the cell loss priority (e.g., violation cells)
- in the event of congestion, drop tagged cells
before dropping any untagged cells
- not to be confused with “delay priority”
(i.e., the service classes)

45
 Summary

 ATM has much confusing terminology, and

an endless (growing) list of TLA’s (three
letter acronyms)
 Get used to it!
 Soon you will be speaking it too!!!

46
ATM Cell Networking

47
 Definition

 The term “cell networking” means packet

switching, but with fixed size packets
(called cells)
 Contrast with “frame relay” which uses
variable size packets
 ATM is an example of cell networking with
53-byte cells

48
 Rationale for Cell Networks

 Simpler interfaces
 Simpler buffering
 Simpler switches
- slotted system
- synchronous stages
 Better control of delay
 Better control of delay jitter

49
Delay Example

50
 Delay Example (Cont’d)

Job 1, Size 10,

arrives at time 0

+ Job 2, Size 2,
arrives at time 3 Server
=
Depart
Job 1 departs at time 10
Job 2 departs at time 12
51
 Delay Example (Cont’d)

 In the Internet, for example, a small packet

(e.g., a telnet packet) that happens to get
stuck behind a large packet (e.g., an ftp
packet) can experience a large delay
 FIFO service, non-preemptive

52
 Delay Example (Cont’d)

Job 1, Size 10,

arrives at time 0

+ Job 2, Size 2,
arrives at time 3 Server
=
Depart
Job 1 departs at time 12 (assumes jobs
Job 2 departs at time 7 are equal priority)
53
 Delay Example (Cont’d)

Job 1, Size 10,

arrives at time 0

+ Job 2, Size 2,
arrives at time 3 Server
=
Depart
Job 1 departs at time 12 (assumes job 2
Job 2 departs at time 5 is higher priority)
54
 Advantages of Cells

 High priority or delay-sensitive traffic will

likely spend less time “stuck behind” other
traffic
 The smaller the cell, the better
 Lower mean delay, and lower variation of
delay
 Easier to provide performance guarantees to
integrated traffic
55
 Summary

 In addition to the ease of implementation

considerations, cell based networks offer a
better framework for providing delay
guarantees on integrated traffic flows (e.g.,
data, voice, video)
 That is why ATM uses cells

56
 Why 53 bytes?

 The smaller the cell, the better (in terms

of delay guarantees)
 Need to design for traffic with the most
stringent delay requirements
 Considerations for voice traffic were an
overriding concern

57
 Why 53 bytes? (Cont’d)

 The ATM cell size was chosen by the

CCITT international standards committee
(now called ITU)
 Influenced by voice traffic requirements and
existing telco equipment in place at the time
(e.g., echo cancellation)

58
 Why 53 bytes? (Cont’d)

 European community wanted 32 bytes of

data per ATM cell
 American community wanted 64
 Result: compromise!
- (32 + 64) / 2 = 48
- thus, 48 bytes of data per ATM cell
 Both sides equally (un)happy

59
 Why 53 bytes? (Cont’d)

 European community wanted 4 bytes of

header per ATM cell
 American community wanted 6
 Result: compromise!
- (4 + 6) / 2 = 5
- thus, 5 bytes of header per ATM cell
 48 + 5 = 53 bytes per ATM cell

60
 Why 53 bytes? (Cont’d)

 Equally inefficient for all types of traffic

(data, voice, video)
- data networks want big packets
- ATM overhead is 5/53 = 10% (too high!)
- voice networks want small(er) packets
- 48 bytes @ 64 kbps = 6 msec
- video probably wants big(ger) packets

61
 Summary

 53 bytes is now the international standard

for ATM cell size
 “Only a standards committee could come up
with a packet size that is a prime!’’
(Raj Jain, 1993)

62
 The ATM Airport:
VPI / VCI Switching Explained

63
 Introduction

 ATM terminology is confusing

(e.g., Virtual Paths, Virtual Channels, VPI’s,
VCI’s, VPC’s, VCC’s, PVC’s...)
 One way to explain these terms is with the
use of a simple analogy: airline travel

64
 The ATM Airport Analogy

 Flight number on a  Virtual Path

specific airline Identifier (VPI)
- e.g., AC 1290 - e.g., VPI = 23

65
 The ATM Airport Analogy

 Flight number on a  Virtual Path

specific airline Identifier (VPI)
- e.g., AC 1290 - e.g., VPI = 23
 Seat assignment on a  Virtual Channel
specific flight Identifier (VCI)
- e.g., 22A - e.g., VCI = 305

66
 The ATM Airport Analogy

 Virtual Channel Connection:

- an end-to-end concatenation of flights and seat
assignments that get you (an individual traffic
flow) to your actual destination
- Example: from Saskatoon to Toronto requires
going Saskatoon-Regina-Winnipeg-Toronto
 Note that VPI’s and VCI’s are only locally
significant (per hop basis)

67
 The ATM Airport Analogy

 Provides a two-level addressing scheme that

uniquely identifies each “cell” (passenger)
on a per-hop basis
 All VCI’s represent individual traffic flows
 VPI is a “bundle” of VCI’s all heading in
the same direction

68
 The ATM Airport Analogy

 All VCI’s on that VPI receive the same

“grade of service” in some sense (e.g., food,
cost, arrival time, bumpy flight, crash, etc.)
 There might be other VPI’s between the
same two points that offer different quality
of service (e.g., other airlines, other flights
at different times of day)

69
 The ATM Airport Analogy

 Airlines (and air traffic controllers) only

deal with VPI’s (i.e., flights) when doing
scheduling, takeoff, landing, routing,
provisioning, etc (not individual cells)
 Airlines can add or remove flights (VPI’s)
on a medium to long term basis, but
individual passengers (VCI’s) can come and
go on a fairly short term basis

70
 The ATM Airport (Cont’d)

 Airport terminal  ATM switch

- Lots of flights and - Lots of cells with
passengers coming in VPI’s and VCI’s
and going out coming in, going out
- Main goal is to make - Main goal is to make
sure that passengers sure that cells coming
coming in on flights in on input ports are
are sent out on the right switched onto the
outgoing flights correct output ports

71
 The ATM Airport (Cont’d)

 An incoming passenger arrives on seat A

of flight B at gate C, and wants to depart on
seat D of flight E at gate F
 Changing flights and seat: VP/VC switch
 Changing seats, but not flight: VC switch
 Changing flight, but not seat: VP switch
 Same flight, same seat: no switch!

72
 Strengths of the Analogy
 Provides nice explanation for VPI’s as
“bundles of VCI’s” heading to same place
- Network management, routing, resource
allocation deals with VPI’s, not VCI’s
 Emphasizes locally significant nature of
VPI and VCI, but end-to-end notion of
virtual channels and virtual paths
 Explains ATM switching in its role as “label
multiplexing”
73
 Weaknesses of the Analogy

 VCI’s in ATM actually correspond to a

traffic flow (stream of cells) not just an
individual cell
 Cells are sent sequentially on ATM links,
not in batches like airline flights
 QOS notions of cell loss, cell delay, and cell
delay variation don’t really fit analogy well
 Does not explain why baggage gets lost!!!

74
 Summary

 The “ATM Airport” offers a clever analogy

for explaining and understanding the role
of VPI’s and VCI’s in ATM networks
 VPI’s correspond to flights
 VCI’s correspond to individual traffic flows
 Airports are the switching hubs that get you
to your proper destination

75
 ATM Adaptation Layer

 The ATM Adaptation Layer (AAL) defines

the rules for breaking up a higher level
protocol data unit (PDU) into ATM cells for
transmission on the network, and for
reassembling the PDU at the other end
 Segmentation and reassembly (SAR)
 Convergence sublayer: adds sequencing and
control information to aid SAR process
76
Example: ATM Adaptation Layer

TCP TCP
IP IP
AAL AAL
ATM ATM ATM
NETWORK

77
 AAL Protocols

 There are several different proposed AAL

protocols for different traffic types
 Example: AAL 1
- for real time or continuous bit rate services,
such as video traffic
- requires 1 byte of overhead per ATM cell
(i.e., within the 48 byte payload)
- 1 bit CSI, 3 bit sequence count, 3 bit CRC, and
1 bit parity
78
 AAL Protocols (Cont’d)

 Example: AAL 3/4

- an adaptation layer for connectionless or
connection-oriented data traffic
- 4 bytes overhead per ATM cell
- 16 bit header (2 bit type, 4 bit seq, 10 bit MID)
- 16 bit trailer (6 bit length, 10 bit CRC)
- very high overhead!!!

79
 Adaptation Layer : A detailed view
Higher PDU
layer

80
 Adaptation Layer : A detailed view
Higher PDU
layer

CS
CS-header PDU Pad CS-trailer
layer

81
 Adaptation Layer : A detailed view
Higher PDU
layer

CS
CS-header PDU Pad CS-trailer
layer

SAR SAR-PDUSAR-PDUSAR-PDU SAR-PDUSAR-PDUSAR-PDU

layer header payload trailer header payload trailer

82
 Adaptation Layer : A detailed view
Higher PDU
layer

CS
CS-header PDU Pad CS-trailer
layer

SAR SAR-PDUSAR-PDUSAR-PDU SAR-PDUSAR-PDUSAR-PDU

layer header payload trailer header payload trailer

ATM Cell Cell Cell Cell

layer header payload header payload
83
 AAL Protocols (Cont’d)

 Example: AAL 5
- an adaptation layer for data traffic designed by
data neworking researchers
- Simple and Efficient Adaptation Layer (SEAL)
- much lower overhead
- 8 byte trailer per PDU (i.e., last cell of a PDU)
- 1 bit of header in last cell of a PDU
- 48 bytes of data in each ATM cell

84
 AAL 5

User data

5 byte 48 bytes flag 8 byte

header of data trailer

85
 AAL 5

Control Length CRC

User data Pad
field field 32

0-47 bytes 2 bytes 2 bytes 4 bytes

86
 Summary

 The ATM Adaptation Layer (AAL) defines

the rules for breaking up a higher level
protocol data unit (PDU) into ATM cells for
transmission on the network, and for
reassembling the PDU at the other end
 Several AAL’s to choose from
 AAL5 is the best for data traffic

87
An Overview of
ATM Signalling

88
 Introduction

 ATM is a connection-oriented protocol,

which means that virtual channels must be
set up before any data cells can be sent on
the channel
 Connection setup is done using a signalling
protocol

89
 ATM Signalling Protocol

 Protocolconsists of two parts

 User-Network Interface (UNI)
- defines how hosts talk to switches
 Network-Network Interface (NNI)
- defines how switches talk to other switches
 Cell formats slightly different

90
ATM UNI Cell Format

GFC VPI
VPI VCI
VCI
VCI PT RES CLP
HEC

Payload
(48 bytes)

91
ATM NNI Cell Format

VPI VPI
VPI VCI
VCI
VCI PT RES CLP
HEC

Payload
(48 bytes)

92
ATM NNI Cell Format

VPI VPI
VPI VCI
VCI
VCI PT RES CLP
HEC

Payload
(48 bytes)

93
 Basic Signalling Operation

 Connection requests proceed hop-by-

hop through the switches of the
network en route to destination
 Switches perform Call Admission
Control (CAC) based on traffic
descriptor, QOS requirements, and
available resources at that switch

94
 Basic Operation (Cont’d)

 If connection is acceptable, then request is

forwarded on, otherwise “reject” is returned
 If destination accepts connection, then
“accept” is returned
 VPI and VCI assigned

95
Example: ATM Signalling

ATM
Network

96
Example: ATM Signalling

ATM Switch
97
Example: ATM Signalling

UNI

98
Example: ATM Signalling

NNI
UNI

99
Example: ATM Signalling

NNI UNI
UNI

100
Example 1: ATM Signalling

CR

CR: Connection Request

101
Example 1: ATM Signalling

CR

CR: Connection Request

102
Example 1: ATM Signalling

CR

CR: Connection Request

103
Example 1: ATM Signalling

CR

CR: Connection Request

104
Example 1: ATM Signalling

CR

CR: Connection Request

105
Example 1: ATM Signalling

OK
(VCI = 12)

CR

CR: Connection Request

OK : Connection Accept
106
Example 1: ATM Signalling

OK
(VCI = 12)

CR
(VCI = 4)

CR: Connection Request

OK : Connection Accept
107
Example 1: ATM Signalling

OK
(VCI = 12)

CR
(VCI = 104)
(VCI = 4)

CR: Connection Request

OK : Connection Accept
108
Example 1: ATM Signalling

OK
(VCI = 12)

CR
(VCI = 104)
(VCI = 4)

(VCI = 4)

CR: Connection Request

OK : Connection Accept
109
Example 1: ATM Signalling

OK
(VCI = 12)

CR
(VCI = 104)

(VCI = 4)
OK
(VCI = 77) (VCI = 4)

CR: Connection Request

OK : Connection Accept
110
Example 2: ATM Signalling

CR

CR: Connection Request

111
Example 2: ATM Signalling

CR

NO

CR: Connection Request

NO: Connection
112 Reject
Example 3: ATM Signalling

CR

CR: Connection Request

113
Example 3: ATM Signalling

CR

CR: Connection Request

114
Example 3: ATM Signalling

CR

CR: Connection Request

115
Example 3: ATM Signalling

CR
NO

CR: Connection Request

NO: Connection Reject
116
Example 3: ATM Signalling

CR
NO

CR: Connection Request

NO: Connection Reject
117
Example 3: ATM Signalling

CR
NO
NO

CR: Connection Request

NO: Connection Reject
118
 Summary

 ATM UNI and NNI signalling specifications

define the protocols for connection setup
and teardown between ATM equipment
(Version 4.0 soon)
 Can support point to point and multipoint
connections

119
 Where to get more information

 ATM UNI 3.0 Specification (now obsolete!)

 ATM Forum

120