基于用户行为的流媒体分段缓存算法

Segmentation Caching Algorithm of Streaming Media Based on User Accessing

对网络中流媒体对象的用户访问行为进行了建模分析,根据相对流行度对缓存中的对象进行排序,同时结合指数分段算法,提出了基于相对流行度的流媒体分段缓存算法(RP-S),并使用真实日志记录进行仿真.仿真实验结果表明,与传统流媒体缓存算法相比,该算法具有较高的字节命中率,同时在一定程度上减少了网络延迟.

Analysis of Internet user accessing of streaming media are presented in this paper at first.Streaming media object shows larger size,static content and high accessing probability on Prefix Segment and that is different from Web object.Relative Popularity（RP） is employed in this paper based on the user accessing of streaming media,and objects are ordered by Relative Popularity in the cache.Integrating with the existing exponential segment,an algorithm based on Relative Popularity（RP） named RP-S is employed.Experimental result shows that algorithm presented in this paper has a higher byte hit ratio compared with the departed caching algorithm of streaming media,and also reduces the occupied bandwidth and user latency.