A loss detection Service for Active Reliable Multicast Protocols

Moufida MAIMOUR & C. D. PHAM

INRIA-RESO

RESAM UCB-Lyon – ENS Lyon

INC’02, Plymouth Tuesday, July 16th, 2002

Outline

• Introduction

• The DyRAM protocol

• The active loss detection service

• An active-based reliable multicast architecture

• Some results (analysis, simulation, implementation)

• Conclusion

From unicast…

• Problem Sending same data to many receivers via unicast is inefficient.

Sender

data

datadata

data

Receiver Receiver Receiver

datadata

…to multicast on the Internet.Sender

data

datadata

data

Receiver Receiver Receiver

ProblemSending same data to many receivers via unicast is inefficient.

SolutionUsing multicast is more efficient

Reliable multicast

• At the routing level : IP Multicast provides efficient delivery without any reliability guarantees.

• Many multicast applications require reliability.

• Reliability has to be addressed at a higher level.

Reliable multicast protocols

• End-to-end solutions :

Only the end hosts (the source and/or the receivers) are involved.

• In-network solutions :

Routers are involved in the recovery process.

Active routers-based solutions

What are active routers ?

Active routers are able to perform customized computations on the messages flowing through them.

DyRAM main characteristics

• DyRAM is based on active services (router-assisted).

• the recovery is performed from the receivers (no data cache at the routers)

• A recovery tree is constructed on a per-packet basis via a replier election mechanism.

• Use of NACKs combined with periodic ACKs.

Main Active Services in DyRAM

• NACK suppression

• Subcast of repair packets

• Dynamic replier election

NACKs suppression

NACK4NACK4

NACK4

NACK4data4

NACK4

only one NACK is forwarded to the source

Replier election and subcast

IP multicastIP multicast

IP multicast

DyRAMDyRAM

IP multicast

IP multicast

DyRAMDyRAM

R1

R2R3R4

R5 R7

0

12

1 0

NAK 2,@ NAK 2,@

NAK 2,@

NAK 2 from link2NAK 2 from link1

NAK 2

Repair 2

Repair 2

Repair 2

Repair 2

D0

D1

NAK 2

NAK 2

The active loss detection service

NACK4

NACK4

NACK4

NACK4data4

NACK4

A NACK is sent by the router

The active loss detection implementation

The Track List (TL) structure which maintains for each multicast session,

• lastOrdered : the sequence number of the last received packet in order

• lastReceived : the sequence number of the last received data packet

• lostList : list of not received data packets in between.

The active loss detection implementation (cont.)

• On reception of a data packet with a sequence number seq > TL.lastOrdered+1

• for each lost data packet (TL.lastOrdered < lostseq < seq & lostseq Є TL.lostList),• send a NACK for lostseq toward the source.

• ignore similar NACKs from downstream links for a given period.

Where to place the active routers ?

core networkGbits/s

wireless LAN1Mbits/s, 10MBits/s

PSTN

10Mbits/s

GSM, UMTS

visio-conferencing

ISDNxDSL

100Mbits/s

Server

Location of the loss detection-capable routers

The loss detection service should be located not too far from the source so the corresponding overhead is justified !

core networkGbits rate

sourcesource

The active router associated to the source can perform early processing on packets.

A hierarchy of active routers can be used for processing specific functions at different layers of the hierarchy : NACK suppression, subcast, replier election.

Specialized active routers architecture

Simulation model

Simulation results

p=0.25

#grp: 6…244 receivers/group

DyRAM implementation

• Tamanoir execution environment• Java 1.3.1 and a linux kernel 2.4• A set of receivers and 2 PC-based routers

(Pentium II 400 MHz 512 KB cache 128MB RAM)

• Active processing cost of a • data packet : 20 micro sec• NACK packet : 135 micro sec• repair packet : 123 micro sec

Conclusion & future work

• Reliability on large-scale multicast session is difficult. Active services at the edges can provide efficient solutions for reducing implosion, recovery delays and exposure problems and so achieving scalability.

• Optimizing the replier election based on an estimation of the receivers power (by means of BW, delay …)

• A congestion control is currently under evaluation and will be integrated into DyRAM in the near future.