P2P流媒体系统中基于声望的节点故障检测方法

Reputation-based method to detect failed peers in P2P streaming media system

针对P2P流媒体系统中由于复杂的网络环境与节点的高度动态性而导致系统整体可用性偏低的问题,进行深入分析,提出了一种基于声望的节点故障检测方法,该方法能够根据节点的反馈构建目标节点的声望,从而有效地检测节点故障。理论分析和仿真结果表明:该方法能够充分利用网络节点的资源,有效地减轻了索引服务器的负载,准确地检测了系统中的故障节点。

Due to complex network environment and high dynamics of peers,it＇s difficult for P2P streaming system to gain high availability.This paper proposes a novel reputation-based method to detect failed peers.The method can construct the reputation of target peers according to feedbacks of peers,hence detect failed peers efficiently.Theoretical analysis and the results of simulation indicate that the method can fully take advantage of resource of peers to alleviate the load on index servers effectively,meantime detect failed peers correctly.