支持内存数据库索引缓存优化的CST树的设计与实现  

Design and Implementation of Cache Sensitive T-Tree Supporting Index Cache Optimization in Main Memory Database

作　　者：唐军 卢正新[2] 

机构地区：[1]广州南方电力集团科技发展有限公司,广州510245 [2]华中科技大学通信与软件交换研究中心,武汉430074

出　　处：《计算机与数字工程》2010年第1期173-176,共4页Computer & Digital Engineering

摘　　要：针对目前内存数据库中索引缓存失配的问题,在分析了现有内存数据库索引结构基础上,提出了一种缓存敏感T树(CST树)的索引数据结构,详细数据结构描述和操作算法也已给出。通过CST树的缓存次数分析和进行查询、插入等操作性能测试,结果表明CST树能有效减少缓存敏感次数,并且在数据量较小时,CST树的插入、删除速度比T树略慢,而查询速度比T树要快。在数据量较大时,CST树的插入、删除、查询效率都比T树要高。Aiming at the cache mismatch problem of index in main memory database, a new index data structure named Cache sensitive T-tree is introduced by analyzing present index data structure and detailed description on data structure and operation algorithm are also given. Finally, the properties of Cache sensitive T-tree is tested in this paper. According to the test data, it is concluded that Cache sensitive T-tree could reduce the times of cache sensitive effectively and when the amount of data is small, the inserting and deleting speed of Cache sensitive T-tree is a bit slower than that of T-tree while the query speed of Cache sensitive T-tree is quicker than that of T-tree. When the amount of data is large, the inserting, deleting, and query speed of Cache sensitive T-tree is quicker than that of T-tree.

关 键 词：内存数据库 T树 索引 缓存优化 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]