Research On P2P Streaming Media Incentive Mechanism

In recent years, with the development of broadband networks, media streaming applications are becoming more and more pervasive in our daily life. The traditional C/S based streaming systems are hard to adapt to large-scale applications due to the server bottleneck. With P2P technology, each peer in the network acts as both a service consumer and provider. By efficiently utilizing the resources at the network edge, server bandwidth can be significantly reduced.. However, the performance of P2P network highly depends on the cooperation among peers. The existence of free-riders will hurt the performance of P2P networks. Bringing incentive mechanisms into P2P systems can not only inspire peers to contribute their resource to others, but also punish free-riders and achieve better system performance.In this thesis, we first survey various technologies and theories on P2P streaming applications. By comparing different P2P network topologies and streaming models, we identify the main shortcomings of existing approaches. , Particularly, we focus on the research on incentive mechanisms in P2P streaming systems. We proposed a Credit-line based Layered Taxation (CLT) incentive mechanism, which can be well applied in P2P live streaming systems. By introducing Credit-line mechanism, we can not only solve excessive dependency on central servers in micro-payment model, but also avoid the reputation accumulation problem and trade limitation in the BitTorrent tit-for-tat model. In addition, layered topology and taxation strategy can inspire peers to contribute their bandwidth, which maximizes both individual utility and system utility.. We describe the above incentive mechanism in details, including network topology, taxation strategy, operation mechanism, collusion precaution and so on. Finally, we conduct extensive simulations on the platform of p2pstrmsim. Our simulation results indicate that, with our proposed incentive mechanism, cooperative peers can gain much better performance than uncooperative peers on playback delay, playback quality and etc. Free-riders are finally kicked out of the P2P network.