Research And Implementation Of Cache Technology In Grid Database

With the rapid development of grid technology, computational grid and data grid concept have appeared for data applications. Database technology plays a very important role in grid application. But the existing distributed database systems are based on a unified view or common structures. It is worth studying the issues that how to establish the caching mechanism and provide service to for query processing using data cache efficiently under dynamic grid environment.In order to meet the new needs with the constant development of the grid, Grid database becomes a new research theme combining the technology of grid and database, which aims to integrate the distributed and heterogeneous database. As one of the core technologies of grid database, traditional query processing technology can not adapt to dynamic grid environment based on a single technology, caching technology becomes a major concern in the field of grid database technology. And the research involves the issues about how to set up a data cache and further use the data in the cache for query processing service in grid environment.The main contents and innovations of this thesis are as follows:⑴In this thesis, the grid database application environment and application requirements are described. And a NHGridDB database architecture based on OGSA prototype is proposed, on the basis of which a Mobile Agent technology based query processing architecture is designed.⑵Orienting the dynamic properties of grid database running environment, a query processing framework based on cache is presented. In addition, a query processing cost model is addressed and further a cache-based query processing algorithm is designed and implemented in the Prototype system NHGridDB. The proposed algorithm detects task load of each grid node as well as the current status, and in accordance with which schedules the task to adapt to grid environment changing.⑶Cache mechanism is designed and achieved in NHGridDB, which includes the creation algorithm, replacement algorithm, and consistency maintenance strategy. It runs data synchronization at an appropriate time and adjusts query execution according to load of grid nodes.