Top Banner
Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia University of South Australia
45

Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

Mar 31, 2015

Download

Documents

Crystal Braford
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

Modelling and Analysis of the CES Protocol of H.245

Lin Liu and Jonathan BillingtonComputer Systems Engineering

CentreUniversity of South Australia

University of South Australia

Page 2: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

Background• Part of an ongoing project on modelling and

analysing Internet multimedia protocols with CPNs, which is focused on ITU-T recommendation H.323.

• H.323 is a key standard for multimedia communications over packet-based networks, e.g. the Internet.

• The CES protocol is a sub-protocol of ITU-T recommendation H.245, which is one of the core protocols of the H.323 standard.

UniSAUniSA

Page 3: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

Outline

• The CES protocol• Modelling and analysis results of

• The CES protocol with a reliable transport

medium

• The CES protocol with an unreliable

transport medium

• Conclusion and future work

UniSAUniSA

Page 4: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The CES Protocol

The Capability Exchange Signalling (CES) protocol is used by a communication party to inform its peer of its multimedia receive and transmit capabilities.

OK, John. Thank you.

Hi Mary, I can only transmit and receive audio.

UniSAUniSA

Page 5: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

Primitives and messages in the CES entities Primitives and messages in the CES entities

Page 6: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The SDL diagrams of the CES protocolThe SDL diagrams of the CES protocol

Page 7: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

Example MSCs of the CES protocol Example MSCs of the CES protocol

Page 8: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

Example MSCs of the CES protocol Example MSCs of the CES protocol

Page 9: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

Example MSCs of the CES protocol Example MSCs of the CES protocol

Page 10: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The CES Protocol with Reliable Transport Medium

• The CES protocol is designed to be independent of the underlying medium.

• When the CES protocol is applied to an H.323 system, it is required by H.323 to operate over a reliable transport layer.

• A reliable transport layer must preserve sequence, be error-free, and provide flow controlled transmission of messages.

UniSAUniSA

Page 11: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

• Basically, each SDL state transition is modelled Basically, each SDL state transition is modelled as a CPN transition.as a CPN transition.

• Each CES primitive is treated as an atomic Each CES primitive is treated as an atomic event, so when two CES primitives are included event, so when two CES primitives are included in a single SDL state transition, they are in a single SDL state transition, they are modelled as two separate CPN transitions, and a modelled as two separate CPN transitions, and a temporary state is introduced.temporary state is introduced.

Modelling SDL state transitions with CPNsModelling SDL state transitions with CPNs

Page 12: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The Incoming CESE SDLThe Incoming CESE SDL

Page 13: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

Modelling the CES messagesModelling the CES messages

Page 14: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The CPN modelThe CPN model

UniSAUniSA

Page 15: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The declaration nodeThe declaration node

UniSAUniSA

Page 16: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The hierarchical CPN model (1)The hierarchical CPN model (1)

UniSAUniSA

Page 17: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The top level pageThe top level page

Page 18: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The Outgoing_CESE subpageThe Outgoing_CESE subpage

Page 19: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The Incoming_CESE subpageThe Incoming_CESE subpage

Page 20: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The hierarchical CPN model (2)The hierarchical CPN model (2)

Page 21: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The top level pageThe top level page

Page 22: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The Outgoing_Send subpageThe Outgoing_Send subpage

Page 23: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The Outgoing_Receive subpageThe Outgoing_Receive subpage

Page 24: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The Incoming_Receive subpageThe Incoming_Receive subpage

Page 25: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The Incoming_send subpageThe Incoming_send subpage

Page 26: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

• This original model has an infinite state space.This original model has an infinite state space.• We investigate the following two cases:We investigate the following two cases:

• Model AModel A • The queue length of places forChannel and The queue length of places forChannel and

revChannel is set to 3;revChannel is set to 3;• Modulo 2 instead of modulo 256 arithmetic.Modulo 2 instead of modulo 256 arithmetic.

• Model BModel B• The number of times that transition The number of times that transition

TRANSFERreq occurs is limited to 2;TRANSFERreq occurs is limited to 2;• Modulo 2 instead of modulo 256 arithmetic.Modulo 2 instead of modulo 256 arithmetic.

UniSAUniSA

Analysis restrictions Analysis restrictions

Page 27: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The state space for Model AThe state space for Model A

!!

UniSAUniSA

Page 28: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The Occ graph (part)The Occ graph (part) of Model Aof Model A

UniSAUniSA

Page 29: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

The CES protocol can fail if wrapping of sequence numbers happens when there are n outstanding acknowledgments for modulo n arithmetic.

Page 30: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

• Two transitions should be dead when the transport medium is reliable, but they are not.

• The CES protocol can fail if wrapping of sequence numbers happens when there are n outstanding acknowledgments for modulo n arithmetic.

• This protocol could work properly • if there are n outstanding acknowledgments, for

modulo n arithmetic, then do not send a TerminalCapabilitySet message,

• set timer and report problem instead.

Conclusion for Model AConclusion for Model A

UniSAUniSA

Page 31: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The state space report for Model BThe state space report for Model B

UniSAUniSA

Page 32: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The full Occ graph The full Occ graph of Model Bof Model B

UniSAUniSA

Page 33: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

Desired terminal Desired terminal state and state and successful successful capability capability exchangesexchanges

Page 34: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

Proper termination of Proper termination of exchanges and exchanges and discard of expired discard of expired messagesmessages

UniSAUniSA

Page 35: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

• Transitions REJECTindUin and TRANSFERindT are dead transitions under the assumption of Model B.

• When the wrapping of sequence numbers (modulo 2) while there are 2 outstanding acknowledgments (i.e. not yet received by the outgoing CESE) does not occur, this protocol can carry out the capability exchange properly.

• An induction proof is needed to generalise these conclusions for arbitrary modulo arithmetic.

Conclusion for Model BConclusion for Model B

UniSAUniSA

Page 36: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The CES Protocol with Unreliable Transport Medium

• An unreliable transport layer may have message loss, duplication, and the ordering of the messages sent may not be preserved.

• We assume that each message can only be duplicated once at most.

UniSAUniSA

Page 37: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The CPN modelThe CPN model

UniSAUniSA

Page 38: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

The state space report The state space report •Counter with limit 2Counter with limit 2•Modulo 2 arithmetic

UniSAUniSA

Page 39: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

REJECTindUin & REJECTindUin & TRANSFERindT TRANSFERindT can occurcan occur

Page 40: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

Proper terminal Proper terminal states and states and successful capability successful capability exchangesexchanges

Page 41: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA

Inefficient behaviour Inefficient behaviour

Page 42: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

• When the transport medium is unreliable, this protocol can work, but it may be inefficient.

Conclusion for the Unreliable Medium CaseConclusion for the Unreliable Medium Case

UniSAUniSA

Page 43: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

Conclusion • Complete CPN models of the CES protocol have

been created, providing a rigorous specification of the CES protocol.

• No matter whether the underlying medium is reliable or not, the CES protocol can fail if wrapping of sequence numbers happens when there are n outstanding acknowledgments for modulo n arithmetic.

• If the problem with sequence number wrap can be avoided, then when the transport layer is unreliable, the CES protocol may be inefficient.

UniSAUniSA

Page 44: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

• Next step: verify the CES protocol against the service specification.

• We have created a complete and general CES service specification (with CPNs) and currently are working on generating the CES service language from this service model.

• Future work includes investigations of other parts of H.323, based on the experience and methodology gained from this pilot study.

UniSAUniSA

Future Work

Page 45: Modelling and Analysis of the CES Protocol of H.245 Lin Liu and Jonathan Billington Computer Systems Engineering Centre University of South Australia.

UniSAUniSA