Research And Implementation Of Federated Database Based On Xml Metadata

In the large-scale modern enterprises different department uses different information management system to store and use data.On the premise of remaining the independent business and apply style of the legacy application unchanged,how to use a common data transformation platform to connect new applications into it with a unified interface or to share data in the legacy systems is the present research focus about system integration,cross-platform access and usage of heterogeneous data.Based on the analysis of the existing variety of heterogeneous data integration technology,this paper raises a data integration reference architecture which is based on web service and peer-to-peer technology;implements the key parts of DI-PPWS (Data Integration based on Peer-to-Peer and Web Service) which is designed according to the reference architecture;gives a detailed descripton on the architecture and related core technology's research and implementation.The author's research work and practice can be summarized as follows:1) Analysis and research of existing data integration solution;2) Propose DI-PPWS data integration soiution's thought and its technology route;3) Application of web services and peer-to-peer technology in the data integration programs;4) The use of proxy technology to transform the protocols between Web services and peer-to-peer technology;5) Implementation of core modules of the program.The author's main creative points are as follows:1) Propose new type of heterogeneous data integration solution DI-PPWS which is based on Web services and peer-to-peer technology,and give some implementation;2) Using of the central index server is an effective solution to the semantic heterogeneous technical problems in data integration.Experimental results show that the data integration solution raised by this paper is feasible and referential,particularly in the cross-platform and distributed data integration applications has a wide range of practical value.