一种基于流行度和分段适应性的流媒体缓存算法  

A POPULARITY-BASED AND SEGMENT ADAPTABILITY-BASED CACHING ALGORITHM FOR STREAMING MEDIA

作　　者：杨菲菲[1] 陈志云[1] 曾秋梅[1] 

机构地区：[1]华东师范大学计算中心,上海200062

出　　处：《计算机应用与软件》2010年第7期227-229,264,共4页Computer Applications and Software

摘　　要：为提高流媒体代理服务器的缓存效率,提出一种基于流行度和分段适应性的流媒体缓存策略。该策略在主流分段缓存替换算法的基础上充分考虑了用户的访问特性,采取基于片段流行度的分段缓存管理策略,将媒体对象内部两点流行度的因素纳入缓存替换策略,改善了流媒体缓存管理的效率、提高了缓存的命中率。利用实际用户访问数据,将该算法与等长分段的缓存算法和指数分段的缓存算法进行了比较,仿真结果证明该算法可以在获得与这些算法相近的请求延迟率的条件下,取得最高的字节命中率。A popularity-based and segment adaptability-based caching strategy for streaming media was proposed in order to improve the caching efficiency of the streaming media proxy server.The strategy,which is based on mainstream segmented caching replacement algorithm,fully takes users＇ access characteristics into consideration,it uses the fragment popularity-based strategy of the segmented cache management to take the factor of two-point popularity of media object into cache replacement strategy,this improves the efficiency of streaming media in cache management and the hitting rate of the cache.The proposed caching algorithm has been compared with the caching algorithms of isometric segmentation and exponential segmentation utilising actual users＇access data,the results of simulation demonstrate that this scheme can achieve highest byte hitting rate while in similar request delay rate as these algorithms have.

关 键 词：流媒体 代理缓存 片段流行度 预取 

分 类 号：TP368.5[自动化与计算机技术—计算机系统结构]