Dynamically Adjustable Heterogeneous Data Integration Based On Web Service

As the development of computer application and network, the requirement of enterprises for heterogeneous data integration becomes more and more strong, and the object to been integrated not only includes structural data as traditional database system, but also includes semi-structural and no-structural data as XML. People regard the research and implement of heterogeneous as more important.We first introduce the background of heterogeneous data integration and the research statement at home and abroad. Then on the research base of some data model, we bring up an XML-based EntityNode Data Model(XEDM), we analyze the architecture of some data integration system and decide to use the mediator/wrapper pattern, we provide a layer of Web Service on each heterogeneous data source to manipulate the data source. The heterogeneous data integration architecture includes fours level: source schema level, local schema level, global schema level, user schema level. There are three transformations between the layers. The transformation from the source schema level to the local schema level is based on Web Service; they are transformed to the data model XEDM. The transformation from the local schema level to the global schema level is the mergence of different local schemas and the solution of confliction. The transformation from the global schema to the user schema is dynamically adjusted by the user authorization and the real-time network situation. Then we particular introduce the integration engine and the process of data integration, the search engine and the optimization of search process. Last we present the detailed design and implement of the data integration system.The research we have done in heterogeneous data integration is helpful for dynamically adding and deleting the heterogeneous data source. By now, the successes we have achieved include the data model, the architecture, the design and implement of the system. The shortcoming is that it needs people to participate in the process of data schema integration, to identify the semantic equivalent entity and resolve the confliction between schemas. In the last of the paper,we put forward some new concepts and object of research we will do in the future.