Research On Distributed Query Algorithm And Implementation In Database Grid

Distributed query processing technique is a difficult problem both in traditionaldistributed database environment, and in the database grid environment, and Grid also has itsparticularity. Currently, there is little work and result in the research of query processing ingrid. This thesis is based on the existing grid standards and technologies, focused on thedistributed relation join algorithm a to adapt to the distributed grid environment, and design arelatively complete solution to achieve efficient distributed query function on top of thecurrent data integration infrastructure.First, the thesis analyzes the characteristics of the database grid, and then presentsrelation join algorithm in database grid, Symmetrical Same-Ordered Semi-Join. Thisalgorithm can reduce the communication cost and increase parallelism, and it applies tomulti-join. This thesis makes analysis of communication cost and performance experiment ofthis algorithm.Secondly, based on the characteristics of this algorithm, this thesis researches thecorresponding SQL query processing method to generate high performance execution plan ofthe join algorithm.Finally, the thesis designs the model of database grid integrated query system. Thissystem is based on OGSA framework and OGSA-DAI, implements symmetricalsame-ordered semi-join and corresponding query processing method. This thesis describes thesystem architecture, defines the interface and behavior of services, describes the system’sdeployment, and makes a detailed description of the operational mechanism, including serviceregistration, creation, executing process, and workflow model.