Research On Data Integration Technology And Its Application

With the growing range of the Audit of Social Security, the business information needs to be processed has some properties, such as multisource, heterogeneity and magnitude, so how to share information among the heterogeneous distributed data sources has become a burning issue. The data integration technology this thesis researches integrates heterogeneous distributed data sources logically, it creates a virtual database by means of middleware, which provides an uniform query interface for the users, and hides heterogeneous data sources below. The data integration technology may be applied to the query in the Audit of Social Security with distributed business information.The design of data integration system framework and the research of data integration algorithm are the main contents of the thesis. The design of data integration system framework comprises the designs of system architecture, the logical structure of middleware, data persistent layer, components interactivities. The structure design of middleware is the key point, which is made up of grammar analyzer, query rewriter, wrapper, connection pool and result integrator logically. The algorithms of data integration include query rewriting algorithm, and result integration algorithm. Query rewriting algorithm will reformulate global-schema-oriented query into data-source-schema-oriented sub query, it expands the query statements on the base of trifurcate link tree, and maps the nodes of schemas and attributes by category. The result integration algorithm mainly processes result sets in respects of connection operation, group operation and nested query; connection operation sorts the result sets using the technology of preprocessing, group operation grounds on balanced tree, nested query utilizes hash mapping and iterator. Finally, this thesis will test and analyze the algorithms given above.