rStream: Resilient and Optimal Peer-to-Peer Streaming with Rateless Codes

DMCA

rStream: Resilient and Optimal Peer-to-Peer Streaming with Rateless Codes (2008)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Chuan Wu , Baochun Li

Venue:	IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
Citations:	34 - 1 self

BibTeX

@ARTICLE{Wu08rstream:resilient,
    author = {Chuan Wu and Baochun Li},
    title = {rStream: Resilient and Optimal Peer-to-Peer Streaming with Rateless Codes},
    journal = {IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS},
    year = {2008},
    volume = {19},
    pages = {77--92}
}

OpenURL

Abstract

Due to the lack of stability and reliability in peer-topeer networks, multimedia streaming over peer-to-peer networks represents several fundamental engineering challenges. First, multimedia streaming sessions need to be resilient to volatile network dynamics and node departures that are characteristic in peer-to-peer networks. Second, they need to take full advantage of the existing bandwidth capacities, by minimizing the delivery of redundant content and the need for content reconciliation among peers during streaming. Finally, streaming peers need to be optimally selected to construct high-quality streaming topologies, so that end-to-end latencies are taken into consideration. The original contributions of this paper are two-fold. First, we propose to use a recent coding technique, referred to as rateless codes, to code the multimedia bitstreams before they are transmitted over peer-to-peer links. The use of rateless codes eliminates the requirements of content reconciliation, as well as the risks of delivering redundant content over the network. Rateless codes also help the streaming sessions to adapt to volatile network dynamics. Second, we minimize end-to-end latencies in streaming sessions by optimizing towards a latency-related objective in a linear optimization problem, the solution to which can be efficiently derived in a decentralized and iterative fashion. The validity and effectiveness of our new contributions are demonstrated in extensive experiments in emulated realistic peer-to-peer environments with our rStream implementation.

Keyphrases

rateless code optimal peer-to-peer streaming peer-to-peer network redundant content content reconciliation end-to-end latency volatile network dynamic rstream implementation emulated realistic peer-to-peer environment several fundamental engineering challenge bandwidth capacity multimedia streaming recent coding technique full advantage peer-to-peer link peer-topeer network linear optimization problem node departure latency-related objective streaming session multimedia streaming session original contribution extensive experiment new contribution multimedia bitstreams iterative fashion

DMCA