摘要
集群技术在大型数据库应用系统中得到了越来越多的应用。无共享结构的集群易于实现,具有良好的可扩展性。但是目前的数据库集群工具非常少,往往与数据库相关。针对这一问题,该文提出了一种灵活有效的构建数据库集群的方法,研究并实现了并行数据库中间件StarTP。StarTP的基本思想是:屏蔽后端数据库的细节,为应用提供单一的虚拟数据库;通过流水并行加速数据加载;利用数据划分和复制将查询本地化从而实现并行查询。StarTP支持大型数据库集群,具有容错和负载均衡功能。试验结果证明了StarTP的有效性和可扩展性。
Clusters of workstations become more and more popular to large database application systems. A shared-nothing architecture is relatively straightforward to implement and has demonstrated both speedup and scale up to hundreds of processors. But the few tools that exist for clustering databases are often database-specific. Star TP addresses this problem. Star TP is a flexible and efficient middleware for database cluster, It presents a single virtual database to the application through CORBA and does not require applications know the details of backend database. It speeds up data loading through pipelined parallelism. And it implements parallel qnery processing through query localization based on data partitioning and replication, Star TP is open, configurable and extensible to support： large database cluster architectures offering fault tolerance and load balance. Experiment results show the efficiency and scalability of Star TE
出处
《计算机工程》
EI
CAS
CSCD
北大核心
2006年第22期55-57,共3页
Computer Engineering
基金
国家"863"计划基金资助重点项目(2004AA112020
2003AA111020
2003AA115210
2003AA115410)
关键词
外行数据库
中间件
集群
Parallel database
Middleware
Cluster
