基于哈希的快速流媒体数据重组算法(HDRR)  

Hash-based Rotating Right Technique(HDRR) for Streaming Data Reorganization

作　　者：尹杰[1] 张松[1] 

机构地区：[1]中科院计算技术研究所,中科院研究生院北京100080

出　　处：《计算机工程与应用》2005年第17期33-35,100,共4页Computer Engineering and Applications

基　　金：国家自然科学基金资助项目(编号:60473096)

摘　　要：流媒体服务器存储发生扩展时,为了充分利用资源,原有媒体数据必须进行重新分配,即数据重组。数据重组的重组代价是一个关键因素,其中包括计算量和移动数据量两个方面。论文提出了一种数据重组算法,在保证较好负载平衡的基础上有效地减少了重组代价,并给出实验数据及对比分析。Media data distributed on many storage nodes in streaming media server must adapt to scalable storage.When the changes take place,to make full use of the system resource,the media data in the system must be reorganized on the new storage units,namely data reorganization.For the on-line service system,the cost of reorganization is an important factor,which includes the computation time and the number of moved blocks.To lessen cost,the authors have proposed a hash-based rotating right technique,and compared it with other proposed techniques and demonstrated the effectiveness through experimentation.

关 键 词：数据重组 流媒体服务器 数据存储 可扩展性 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]