摘要
AnyQuery是一种适用于网格环境、基于服务的分布式查询处理系统,使用统一的数据格式(可扩展标记语言)来表示和传输数据,所有组成模块以网格服务的形式实现,网格服务之间通过标准的接口进行通信.这种接口和实现分离技术屏蔽了不同数据库及其运行节点的差异,为用户提供一致的虚拟关系数据视图和分布式数据查询支持.在分析和研究已有的并行数据库和分布式数据库查询技术的基础上,AnyQuery提出和实现了一种适用于动态网格环境的分布式查询计划生成和执行算法,该算法能够根据网格环境中各节点及其连接网络的实际情况确定查询计划,支持跨数据库的连接查询.基于Globus3.0开发了AnyQuery的原型系统,性能测试表明该系统具有较好的可扩展性,尤其适合计算密集型的复杂查询.
This paper proposed AnyQuery, a service-based distributed query processing system to solve the problem of accessing and integrating multiple relational databases in the grid. All the modules of the system are implemented as grid services, which communicate with each other through standard interfaces. XML is used as the data representing and exchanging format. The interfaces and implementations are totally separated in this system. AnyQuery provides a uniform virtual data view above multiple relational databases in the grid. Grid users can issue their queries based on the virtual view. A distributed query processing algorithm is presented to deal with dynamically changed grid environment, which can make the query plan according to the varied grid computing and communicating capacities and support join query across multiple relational databases. A prototype is developed and the performance evaluation demonstrates its effectiveness, especially for complicated computing-intensive query.
出处
《小型微型计算机系统》
CSCD
北大核心
2006年第8期1432-1438,共7页
Journal of Chinese Computer Systems
基金
国家杰出青年科学基金项目"基于数据网格的高性能存储环境及其关键技术的研究(60125208)资助
武汉市青年科技晨光计划"网格环境下数据库访问和集成技术研究(20035002016-03)资助
关键词
分布式查询
服务
网格
distributed query
service
grid
