Research On Information Integration Based On Virtual Database

With development of informatization for a few decades, many computer application systems have been built in the enterprises now. From current situations, it has gone into a key period in the process of the enterprise informatization, which has been realized by many enterprises. Some conceptions or plans such as ERP (Enterprise Resource Plan), EAI (Enterprise Application Integration), Digital Campus, were brought forward, too. From the angle of the Information Engineering (IE), our nation is undergoing the conversions of informatization from high expanding period to system integration period, enterprise data environment from application database to subject database. Under such circumstances, a information integration platform, VDBII, which has three layers (data source wrapper layer, integration processing layer and management layer), based on the Information Integration (II) section of the project of Chongqing University's Digital Campus, is designed in this paper. It uses relational data model as its public data model. VDBII aims at providing a platform integrated many data sources that can be accessed by users transparently and also providing a view of relational database. Firstly, the paper analyzes the background, the actuality at home and abroad and the content of the research, then discusses the relative technologies and introduces the theories of information engineering and system integration. Emphatically, the middleware, the wrapper and the EAI have been canvassed in this paper. Those theories and technologies are stabile foundation for information integration. In Chapter Three, the technologies of virtual database (VDB) and the functions and architecture of the VDBII are discussed. The II in this paper is based on the virtual database, and we emphatically discuss the processing and the architecture of the VDB technology. Under the conditions that we have the high-level overview of the VDBII, we emphatically canvass the design of the data source wrapper. All data sources are divided into two kinds: relational data source and XML data source, An algorithm to convert between them and the methods to construct the converting rules have been put forward. In Chapter Four, the integration processing of the information integration is discussed. The integration processing is the core of the platform and is implemented primarily in the second layer of VDBII. In order to decrease the complexity of the pick-up and mapping of schema belonging to the data sources, a keyword dictionary and a data entity distinguishing module are designed, a method to construct the dictionary and an algorithm to distinguishing entities have also been put forward. Besides discussing how to decompose the schema and combine all the data from various data sources, the algorithm to implement them has been brought forward, too. In Chapter Five, the managements and controls of the II are discussed. They guarantee the platform to run safely and effectively. Emphatically, we discuss session management, transaction management and security management. Task management is employed to implement transaction of the integration task in the platform in a specific period.