Study Of Query Strategy Optimization Of Distributed

Distributed database system is the combination of database and computernetwork logic which is a structured data collection. It distributed in different sites ofthe computer networks in physical; however, it belongs to the integration in logical.But the distribution and redundancy of distributed database makes the data queryoperation complicated. Therefore, the important subject is how to improve the queryefficiency of the distributed database in distributed database research field. It namedquery optimization.The goal of distributed query optimization is to reduce the communication costto a minimum and shorten the response time to the shortest, namely to use a minimumcost, in the shortest response time to obtain the data needed. Base on this target, thereare two major query optimization methods: direct connection based queryoptimization algorithm and semi-join based query optimization algorithm. Firstly, thisresearch studies the direct connection query. Then, analyze the construct sitedependence hash partitioning algorithm. The cost of Hash partitioning algorithmrepeating Hash partition is high, contra posing this drawback, this research adopts thethought of first judging then connecting to improve it, which optimizes the Hashpartitioning algorithm. In addition, analysis the process of general semi-join queryoptimization algorithm, Generally, the average amount of data transmission ofsemi-join query optimization algorithm is large, this research adopts the mode of firstanalyzing the connection data and then carrying out compressed transmission, whichimproves the semi-join query optimization algorithm, reduces the average query cost;SDD-1is a classic semi-join query optimization algorithm, but this algorithm has notconsidered the last-time transmission cost. This research adopts the thought of treatingthe last bit transmission rate as the calculation factor while calculating the cost ofsemi-join, which improves the SDD-1algorithm and the overall query efficiency ofSDD-1algorithm. In this research, depend on experimental tests calculation, we havea result that: the improved algorithm improves the efficiency of the distributeddatabase query.In conclusion, this research still have certain limitations about how to improvedalgorithms of this article, cache-based Hash partitioning algorithm, parallelism of SDD-1algorithm and other issues. This will be the direction of future research.