对等流媒体数据调度优化算法

Research on optimization of data scheduling in peer to peer streaming system

对Mesh结构的对等流媒体系统中的数据调度问题进行了研究。进行了优化算法的分析,利用本地多级缓存来提高数据分片在节点间的分发能力,在此基础上通过将分片播放紧急度和稀缺度加权合并来定义数据优先请求级别;分析了影响节点上传能力的多种因素,采用历史信息统计法来估算节点上传能力,提出基于数据优先级和节点能力评估的对等流媒体数据调度算法。仿真结果表明,基于数据优先级和节点能力评估的调度算法在节点扰动性较强的网络环境中能较好保证节点播放的连续性和数据分片上传能力。

The problem of data scheduling in Mesh-pull P2P streaming systems was investigated and the optimization of schedu-ling algorithms were analyzed.A multi-level buffer mechanism was presented to accelerate chunk distributing among peers, which divided the local buffer into receiving buffer,forwarding buffer and playback buffer.Firstly,based on multi-level buffers, the definition of chunk priority was given,according to data emergency degree in local play buffer and scarcity degree in neigh-bor’s forwarding buffer.Secondly,the factors which influence the peer’s upload capacity were analyzed.Then the implementa-tion of algorithm on data scheduling was given based on chunk priority and peer upload capacity in data-driven P2P streaming sys-tems.Simulation results showed that,even in high dynamic circumstances of P2P network,the new scheduling algorithm could provide better playback continuity and upload capacity.