Research On Data Query Optimization In Distributed Database

Since the 70th in 20 centuries middle, with the rapid development of computer network technology and the wide needs of geographically dispersed companies, groups and organizations, distributed database system generated and has been developing. Query is the most commonly operation that has been used in database. Since the birth of distributed database, the study of query optimization in distributed database has never been stopped in more than 30 years. Because of distribution and redundancy of data distributed, the query optimization in distributed database system is more complex than that in centralized database. Therefore, query optimization is an important problem in distributed database system.Distributed database query optimization algorithms can be divided into semi-join-based optimization algorithm and direct-join-based optimization algorithm for different network environments or query optimization goals. In this paper, we first introduce the concepts of distributed database system and basic knowledge of query optimization. Secondly, we studied several commonly used optimization algorithms and improve the ability of two algorithms.In core of this paper, firstly we describe the performance and default of traditional semi-join algorithm and propose an improved algorithm. This improved algorithm can reduce the cost of network transmission and improve the performance of semi-join algorithm. Secondly, we improved Partition algorithm. In the improved algorithm, we propose a method to divide query graph. This method can divide the query graph into multiple sub-query graphs. Through this method, the parallelism of query can be improved. In the sub-query graphs, Partition algorithm can execute independently and make full use of its advantages. So the overall query execution speed can be raised. Finally, we verified the performance of these two improved algorithm by experiments.Finally we review entire work. We point the insufficient of our work and indicate the goal for the future research.