Research And Implementation Of Heterogeneous Data Share Model Preserving Data Dependency

As techniques of computer network and database progress rapidly, the degree of enterprise informatization becomes higher accompany with frequent data exchange between enterprises. So it's necessary to integrate the data of application systems in order to make the share and use of data resource more effective. Nowadays there's still no universal mechanism for storing data in different enterprise application systems, so it's hard to exchange data effectively. Furthermore, data resources of enterprise distribute widely in different networks, which increases the difficulties of accessing data. Thus, how to exchange data effectively and safely between heterogeneous data resources in distributed environment is critical during enterprise's data integration.As a brief version of standard markup language, XML is suitable to be used as middle carrier during heterogeneous data exchange. Researches about transformation of heterogeneous data schema focus on structure and semantic constraints, which neglects data dependency information. This paper extends inference roles of functional dependency based on XML Schema after a thorough study, and proves the correctness of extended rules. Furthermore, two algorithms are designed in this paper. One is MFDS designed to get minimum functional dependency set, and the other is OFDS designed to optimize minimum functional dependency set. In addition, the rules based on XML Schema to store functional dependency are also established in this paper.After detailed analysis of the method transforming information of semantic constraint and functional dependency information between XML and relational schema, this paper designs and implements the transformation algorithm which can keep functional dependency and semantic constraints between XML and relational data. Furthermore, a data share model based on XML and WebLogic JMS is presented. In this model communication and data transfer between application systems are realized through the message processing mechanism of WebLogic JMS. Shared data are stored in central database in form of XML document and presented to users through message interface, which shields the physical character such as position and type.This paper presents the rules batch downloading of heterogeneous data based on push and draw modes, and gives transaction data management based on authority fields and authority objects. Furthermore, a detailed analysis is given about how to implement distributed transaction management and guarantee security during message transfer. As to data type mapping, data in this model are mapped according to different data precision level so as to decrease the error brought during type transfer. Finally, this paper summarizes the work and presents what should be progressed in further work.