Research And Implement Of Distributed Data Integration Based On Middle Ware

For information systems are often built by stage, the applied systems based on database in governments and enterprises are often distributed and heterogeneous. Thus many information islands came out. With more and more data being accumulated, the problem of information islands became more and more serious. With the development of internet and network technology, more and more information systems want to integrate all of their distributed parts and share data with others. This is the reason of 'data integration'.For the reason of history, the data has been involved in a system often differs in location and structures. Even if the data is same kind in structure, it often differs in storing or visiting models. So it is the main problem to exchange data between heterogeneous systems.There are always many kinds of relative techniques, such as federal database and special data transfer tools. In recent years some new techniques have also been used to integrate data, like XML, P2P, Web Services etc. But these methods are often less of flexibility, extensibility or generality. Howerver, the techonology of middleware can overcome these shortcomings easily .Middleware came into birth with the development of software industrialization, who aims to improve the efficiency and quality of software. The use of middleware in SE should be called as a revolution, which makes the software system could be finished by means of 'assembly parts'. The use of middleware in data integration can gets more performance with less cost. By the construction of a middle layer between data sources and application, an old system can be rebuilt rapidly and cheaply. And the new system will get lots of flexibilities and extensibilities. So it's a good way to integrate distributed and heterogeneous data sources.In this article the data integration and middleware technologies are introduced. A mode] for distributed data integration based on middleware is raised and validated by a real case named HAFSS (House Accumulation Fund Supervising System).Data integration is always a difficult problem in Database field. The researchof it is very practical and valuable. The model reffered in this article pays attention to the practicability and prospect. It has high value when different existing systems are integrated.