Research On Data Query Optimization Algorithm Of Distributed Database

This paper introduces the basic concept of distributed database system, and describes the query process of distributed query.This paper give emphasis to the research on the regular optimization algorithms of distributed query,This paper will lucubrate optimization of queries in distributed database where relations are hash partitioned and bring forward a new heuristic algorithm—a hash partitioning equijoin optimization heuristic algorithm based on Backtracking. This optimized algorithm using the concept of backtrack can divide the query graph into many query blocks by making a depth-first search for the query graph. Then it will determine an effective sequence of join operations for some chosen query blocks by taking advantage of Kruskal heuristic algorithm. When a round depth-first search is terminated, the algorithm will reconstruct a new query graph by merging all query blocks divided from former query graph. The next step this algorithm will take is to recursively make a depth-first search for the already rebuilt query graph until the query graph can not be further divided or all query blocks are unexceptionally composed of only one edge. In the end of this paper, the algorithm is realized by experiment. The result of the experiment is shown that the cost of join operations reqired for distributed queries is more cheaper in contrast with the traditional method, such as Kruskal heuristic algorithm.