摘要： 在基于有服务器结构对等网络环境的点播系统中，针对客户端需要缓存部分流媒体数据的问题，提出一种新型预测双缓存模型，给出相应的缓存替换算法。仿真实验表明，该模型在用户数目增多时，客户端缓存的命中率大幅提高，服务器的负载得到改善，显著降低了VCR操作时所引起的等待延迟。

关键词: 对等网络, 流媒体, 预测缓存, 点播

Abstract: In the Video on Demand(VoD) system that under the Peer to Peer(P2P) network environment with structure of server and as for the client side’s demand on the cache media streaming data, this paper proposes a new prediction double buffer model and gives the corresponding cache replacement algorithm. The model is validated by the simulation experiment that along with the growing number of users, it can greatly raise the hit ratio of client cache, the load of server is improved, and the model significantly lowers the waiting delay during the operating process of Video Cassette Recorder(VCR).

Key words: Peer to Peer(P2P), streaming media, prediction caching, Video on Demand(VoD)

中图分类号: 

• TP393