基于数据网格的分布式查询优化模型  

Distributed query optimization model based on data grid

作　　者：罗永红[1] 陈特放[1] 张友生[2] 

机构地区：[1]中南大学交通运输工程学院 [2]湖南师范大学物理与信息科学学院,长沙410081

出　　处：《计算机应用》2008年第10期2553-2557,共5页journal of Computer Applications

基　　金：湖南省教育厅科研基金项目(06C126)

摘　　要：智能交通系统(ITS)中多个异构的、地理位置分散的数据源能使用像数据网格这样的分布式计算技术进行集成,这种集成所面临的真正挑战是分布式查询处理引擎的设计和开发。一般而言,分布式查询优化按照以下三个阶段进行:查询所涉及节点的确定,并行执行方案的生成,执行查询的最佳节点选择。由于这三个阶段的相互隔离可能会导致得到的查询方案并不是最佳的,提出了一个新的分布式查询优化模型,该模型集成了查询优化的三个阶段,综合考虑了查询优化各个阶段所涉及的参数,如节点的有效内存、处理速度、数据传输容错能力等。The disparate and geographically distributed data sources in ITS （Intelligent Transportation System） can be integrated using distributed computing technologies such as data grid. The real challenge involved in such data integration is in the design and development of the distributed query processing engine. Generally speaking, distributed query processing and optimization are carried out in three distinct phases： determination of single node plan, generation of parallel plan, and optimal node selection for plan execution. As considering the three phases in isolation leads to sub-optimal plans, a new distributed query optimization model was proposed to integrate all the three phases of the query optimization and take all the parameters in the query optimization into account including effective memory, processing speed, tolerance ability of data transportation, etc.

关 键 词：智能交通系统 数据网格 分布式查询 优化模型 

分 类 号：TP393.11[自动化与计算机技术—计算机应用技术]