Research Of Data Partition And Query Optimization Based On Database Cluster

In the development of enterprise informatization, it usually adopt a centralized database system to store business data at the beginning due to the small scale of business and data. But with the expansion of business scale, data quantity will be increasing, the centralized database system would soon become the bottleneck of enterprise information systems, some problems such as low querying speed will happen. In my course of the internship, the host company also encountered this problem, it uses a middleware-based MySQL database cluster to solve this problem. In this context, this paper studied data partition and query processing on the database cluster system, and on this basis to implement some query optimization.In a database cluster environment, generally need to divide the data and then disperse it to the back-end servers, such treatment could bring many benefits, but also introduces some new problems, one of them is query processing. When middleware receives a query request, it must calculates route information according to the division information of the table in the request, and then forwarding the query request to the corresponding back-end database servers according to the route information, each database servers receives the query request and executes it, and then return the results, middleware need to merge the results received from each database servers, and finally returned this merged results to the client, in this process combining with the data partitioning some optimizations of query processing can be done. This article researched the logical of query processing in the data partitioning context, and then introduced some optimization methods for sort and join queries, finally implemented the optimization methods and the corresponding data division algorithms in a middleware-based database cluster system.Finally, in the experimental environment this paper build a database cluster system based on a middleware, by means of this system we tested data partitioning and query optimization, the test results showed that the data partitioning function was properly implemented, and the according query optimization effect is obvious.