论文部分内容阅读
大规模视频点播（ＬＳＶＯＤ）中ｃａｃｈｅ机制是提高系统效率的有效手段，是支持ＶＯＤ实用化的关键技术之一．由于连续媒体的数据量大，使用周期长等特点，传统的ｃａｃｈｅ替换算法不能直接应用于ＬＳＶＯＤ．文中根据ＶＯＤ的特点开发了两种基于访问频率的替换算法：ＬＦＲＵ（ｌｅａｓｔｆｒｅｑｕｅｎｃｙ ａｎｄ ｒｅｃｅｎｔｌｙ ｕｓｅｄ）和ＰＬＦＵ（ｐｅｒｉｏｄ ｌｅａｓｔｆｒｅ－ｑｕｅｎｃｙ ｕｓｅｄ）算法，它们都试图将访问频率大的视频数据保留在ｃａｃｈｅ中．ＬＦＲＵ 算法结合了数据的访问频率和访问时间信息，对访问模式的变化具有一定的适应性．ＰＬＦＵ 算法用周期法和预测法解决了ＬＦＵ 算法中的ｃａｃｈｅ“污染”问题．实验数据显示，新的替换算法比传统的替换算法效率高，更适合大规模的ＶＯＤ系统．
The cache mechanism in large-scale video on demand (LSVOD) is an effective way to improve system efficiency and is one of the key technologies to support VOD practicality. Due to the large amount of data and long service life of continuous media, the traditional cache replacement algorithm can not be directly applied to LSVOD. Based on the characteristics of VOD, two alternative algorithms based on the access frequency are developed: LFRU (least frequency and recently used) and PLFU (period leastfrequenching used) algorithms, all of which attempt to keep the video data with high frequency of access in the cache. LFRU algorithm combines the data access frequency and access time information, and has some adaptability to the change of access mode. The PLFU algorithm solves the cache “pollution” problem in the LFU algorithm using the periodic method and the predictive method. Experimental data show that the new replacement algorithm is more efficient than the traditional replacement algorithm and is more suitable for large-scale VOD system.