Double-ended Cooperation Peer Selection And Rate And Data Distribution For P2P Streaming Media VOD System

With the application of P2P technology and rapid development of broadband network in recent years, the streaming media VOD technology based on the P2P network is becoming one hotspot. The core idea of P2P technology is that it shares the information resources and services through direct interation between peers in the system. The streaming media VOD system based on P2P network can not only overcome the server bottleneck problem in the traditional C/S architecture but also reduce deployment cost and complexity, it shows irreplaceable advantages in the expandability, fault-tolerance and resistance attacks. Peer selection and rate and data distribution are two key technologies in the P2P streaming media VOD system, this paper makes in-depth researches aiming at some issues requiring urgent solution that the two technologies currently have.The main research contents and innovations of this paper are as follows:Firstly, through analyzing and comparing a large number of literatures, this paper summarizes some issues that the existing peer selection algorithms and data scheduling algorithms have. For example, the existing peer selection algorithms not only have single function but also just focus on the implementation of the requesting peers and ignore the selection quality of the candidate peers in the super peers. Some of the existing data scheduling algorithms are based on the traditional C/S mode, and can not be directly applied to the P2P environment; the others are based on the P2P mode, but they study data scheduling among the streaming media data paragraphs, and how to distribute the rate and data for each streaming media data paragraph, they do not study.Secondly, in order to resolve the problems that the existing peer selection algorithms have, a double-ended cooperation peer selection strategy is proposed, which is between super peers and requesting peers. At the super peers, an algorithm of selecting candidate peers based on routing resemblance is proposed, which is topology conglomeration oriented. And at the requesting peers, considering the dynamic and heterogeneity of the system, an algorithm of selecting service peers based on enumeration is proposed, which can guarantee the streaming quality of service and meanwhile improve the robust by adding the thought of dynamic redundancy.Thirdly, in order to resolve the problems that the existing data scheduling algorithms have, a rate and data distribution strategy for each streaming media data paragraph is presented, which is suitable for P2P mode. For each streaming media data paragraph which is waiting for transmission, the strategy conducts allocation based on data packet among different service peers. It is aimed to reduce the total network bandwidth consumptions and meet the need of bandwidth constraint on different network paths at the premise of guaranteeing the playback quality.Finally, a simulation experiment is conducted, which is to test the peer selection strategy and the rate and data distribution strategy proposed in this paper. Simulation results show that the double-ended cooperation peer selection strategy can improve the topology conglomeration, streaming quality of service and robust much better. Beacause the experiment adopts the strategy proposed in this paper to conduct rate and data allocation, the experiment itself fully proves the availability and effectiveness of the rate and data distribution strategy proposed in this paper.