The Research And Application Of Data Modeling And Integration Technology Base On Metadata

With the development of computer technology, especially network technology and information construction, the storage modes of data within an exponential growth pattern are ever-changing. The management of data and business logic aiming at informatization, standardization and platformization has been gradually realized in various industries and trades by data collection, storage, process and transmission. According to different application scenarios and business requirements, departments construct their own information service and management systems which are relatively independent. Due to development stage, developer, purpose and other factors, the information systems always have distinctive logical and physical structures following the inconsistent and incomplete data which make difficult to effective sharing and interaction between systems as well as data exchange within the process of design, management and production. This is known as information islands. Data integration is an feasible and effective way to reduce data redundancy, achieve data sharing and solve information islands problem owe to data exchange between systems.In this paper, we analyse and compare the existing methods of data integration, present a new solution of data modeling and integration based on metadata and build a data integration framework. Firstly, we establish technical and business metadata model according to the sample of shandong university information service platform aiming at information management systems. Secondly, we abstract the platform independent models from business logic and operation of data integration in order to construct business process model, and then obtain the platform specific model including the details of platform information by model transformation. Using semi-structured XML to describe content and structure of models can enhance the reusability and robustness of code and realize the modular management. The application of model driven architecture can improve the systems’expandability and flexibility and reduce the difficulty of operation and coupling between modules.Then, it’s discussed about the implementation of data integration framework based on metadata. Firstly, it’s introduced how to exchange data by interfaces with SOAP, REST Technology. Then, we present three kinds of data synchronization strategy, including instant, timing, trigger&push. Finally, we implement the data process model parser of import and export with MDA and XML.The approach in this paper, data modeling and integration based on metadata, not only has the value of basic theory and critical technique research, but also has certain application worth and practical significance. This work in this paper can provide a feasible, effective and comprehensive solution for data integration research area and university informatization construction and has been applied in Shandong university information service platform successfully.