Study Of The Corba-Based Integrated System Of Heterogeneous Data Sources

In this paper, we study the integration of heterogeneous data sources; discuss the actual problems and the corresponding solutions of it. We propose our CORBA-based heterogeneous data integrated system, analyze the whole function module of integrated system in detail, and discuss the common data model and data integration levels deeply. Then we achieve the functional structure of the heterogeneous data integrated system and apply it into the test problem library system.The architecture of integrated system is three-tier Client/Server; the middleware adopts the mediator-wrapper architecture to complete the main functions of system. The system wrap each local heterogeneous data source with ORB in OIM common data model, and the wrapped CORBA objects provide a uniform access interface and standard form of data exchange. Only the global common model and global data operation can be saw by the user, and the actual data operations are separated into sub-operations according to data sources, then the results are returned to users by integrated system. In addition, we introduce two kinds of roles, the manager of heterogeneous data source and the manager of integrated system, to reduce the difficulties of the implement and solve the problem of authority management. The former completes the login of heterogeneous data source and the definition of shared content; the latter takes charge of combination of the shared data source.On common data model, the definition and interrelated concepts of OIM data model are explained. And then the translation rules between local data model and OEM data model are illustrated. The transformation of Word document aims at the test paper.In the application of system, the archive management module of the Word document is introduced deeply. According to the translating relation of Word document and text file, we use tags to take out the necessary information from semi-structured data to structured data in database, then express these structured data in OIM data model. We implement the transformation from the Word document to the relational database, and prove the validity of our transformation method.