基于可变窗口的流媒体缓存算法研究与实现

Research and Implementation of Streaming Media Caching Algorithm Based on Variable Window

缓存管理是影响流媒体服务器性能的关键因素之一。通过对流媒体访问特性的分析,提出并实现了基于可变窗口的流媒体缓存替换算法CABVW。该算法通过动态调整拆分代价和合并代价的阈值,实现了磁盘带宽和缓存的负载平衡。实验结果表明,该算法与传统主流算法相比明显地增加了流媒体服务器的并发用户数量;算法计算复杂度低,也较易于实现。

Caching management is one of key factors impacting the performance of media streaming server.Based on analyzing the characteristics of streaming media accessing,Caching Algorithm Based on Variable Window(CABVW)is presented in this paper.CABVW keeps the balance between disk I /O bandwidth and the caching system by dynamically adjusting the values of merging -cache cost and splitting -cache cost.The experimental results indicate that CABVW observably increases the number of simultaneous users served by the streaming server;it has a lower computing complexity and its implementation is simple as also.