Research Of Cache-Based Query Optimization Technologies In Swift-Query Tool

Nowadays in a variety of Enterprise Information Systems, especially College Information System, new types of application requirements are emerging. To examplify, users often have unexpected needs to query some specific information. These query requirements bring new challenges to current information systems.This thesis proposes OLAQ(On-line Ad-hoc Query) corresponding to OLTP and OLAP to summarize the requirements and features of swift query application. Then design and develop a new swift query tool named SwiftQuery which is flexible and easy to use. The core idea of this tool is that encapsulating data set which is concerned by business worker by business data object so as to masking technical details of the underlying database schema. The major works of this paper are listed as following:(1) Proposing the concept of BDO(Business Data Object) and use it to compute and encapsulate tuple or attribute of database schema which makes user can use their familiar business data object to express swift query requirements.(2) For swift query requirements, BDO cache model is designed by taking advantage of thoughts of semantic cache. BDO cache abandons the predicate description about result set in traditional semantic cache and describe it by BDO. This way sacrifice part of the memory space to improve the query efficiency.(3) Research on how to express swift query requirements and generate SQL according to combined BDOs. Based on it, proposing a query processing optimization algorithm-bpmTrim(bdo pattern matcher Trim) according to BDO cache, this algorithm can increase efficiency of swift query.(4) Researching the replacement policy of BDO cache, propose ILFU method which considers access frequency, updating frequency, replacement expenses and other elements. This algorithm can improve the performance of BDO cache.(5) Implementing SwiftQuery prototype system based on BDO cache on the basis of above work. Experimental results show that this tool can resolve swift query requirements in practical application very well. Also, the key technologies and other related solutions of this thesis are effective.