Research On The Query Optimization Of Audit Method Based On Audit Platform

Auditing is a process to evaluate finance measure of corporate,national policy implementation by collecting of large amounts of data and analyzing of those data.Yet,auditing is a boring and hard work with low tolerance of mistakes.The application of computer technology greatly releases the auditing officer from the work pressure of auditing,saves the cost of auditing,and improves the accuracy of the auditing.In the era of big data,it is hard to process the vast amounts of audit data timely even with the most advanced computer technologies.Aiming at the processing of auditing,considering with the features of auditing,this project takes fully advantage of the parallelization of computer to speed up the auditing query and analysis,which can improve the audit work effectively.This thesis focuses on the auditing platform of endowment insurance with the considering of the characteristics of the auditing method.Multi-level researchs have been done on performance optimization techniques of the audit database comprehensive,from the host setting to the audit application,from single-threaded query to multicore parallel query,and from a single SQL query to multiple SQL parallel query.Firstly based on the correlation analysis of queried-data,this thesis analyzed the parallelism of target auditing application,and then well designed the submission sequence of SQL to reduce the data competition among queries and improve the reutilization of data cache.Furthermore,with the table partition technology of database and the consideration of data distribution,we researched DB2 database index construction and the methods of parallel query on multicore systems,and analyzed the speedup ratio of multi-core data query;and then analyzed the execution frequency of SQL of the target auditing application.With consideration of the execution frequency of application and the cost from DB2 explain,as current query optimize is lack of the consideration memory cost,we proposed a new run-time SQL query optimization scheduling methods based on the maximum cache reusing.And the resources configuration of Microsoft Windows Server 2003 and DB2 V9.5 are optimized to comprehensively improve the performance of the auditing database,and the testing is done on this target system toverify the correctness and feasibility of the optimization scheme which is proposed in this paper.