P2P直播系统中节点启动过程优化

Startup Process Optimization of P2P Live Streaming System

针对现有P2P直播系统在动态环境下存在的较大启动时延等性能问题,建立节点的启动过程模型,分析影响节点的启动过程性能的关键因素,并从数据调度的角度提出了1种节点状态自适应的启动策略。该策略对节点的启动过程中的请求起始位置的选取、缓冲区的数据填充策略、播放策略3方面进行改进,以提高节点在启动过程中获取数据的效率,减少资源浪费,改进开始播放条件,进而达到提高节点对动态环境的适应性,提高系统的数据分发效率,降低节点的启动时延与播放延时的目的。仿真结果表明,该策略在动态环境下能够降低节点的启动时延与播放延时,提高数据分发效率。

In order to solve the performance problems of existing P2P live streaming system in the dynamic environment,such as considerable startup delay,the node startup process was modeled to analyze the key factors,which had effects on node startup process performance.From the view of data scheduling,a node status adaptive startup strategy（NSAS） was proposed.This strategy improved three major processes of node startup process,which were selecting the startup request point,data scheduling,and playing strategy,so as to reduce the resource waste in startup process,accelerate nodes obtaining chunks and improve start playing conditions.Simulation results showed that this strategy can improve the system adaptability to dynamic environment and data distribution efficiency,and reduce the startup delay and playback lag.