一种Tracker引导的缓存调度方法

Cache Scheduling Method Based on Tracker's Advices

为优化流媒体视频点播系统中客户端对数据的缓存调度,使整个系统的播放质量与资源利用率趋向最优,提出一种Tracker引导的缓存调度方法,Tracker收集缓存服务器的负载信息和网络中的数据请求信息,采用基于流行度的预测方法推测下一阶段网络中需要的副本数,以此为根据对客户端缓存建议请求予以响应.客户端以Tracker的建议和自身在邻域网络内感知到的副本数为依据,使用兼顾全局和局部利益的决策算法进行视频数据的缓存操作.模拟结果表明,Tracker引导的缓存调度方法能够有效的保证网络中客户端节点的播放质量,并在一定程度上降低缓存服务器的负担,使整个系统的缓存调度趋向最优.

To ensure the QoS and reduce the burden of web server in a multimedia streaming system, a cache scheduling method based on trackerg advice is proposed, which guides caching mechanisms in peers by giving advices to peers. All advices are based on the prediction made by tracker server according to the information gathered from peers and proxy servers. On basis of these advices and copy number perceived by themselves, peers can choose appropriate segments of multimedia objects to be cached through decision-making algorithm considering global and local benefit. Simulation results show that this approach can effectively ensure the QoS in local peers and reduce the burden of proxy server in a certain degree and get an optimal solution for the cache scheduling in the whole system.