The Research And Implementation Of Heterogeneous Database Integration System Based On Digital Library

With the development of information technology, digital information resources will gradually become the leading role in the digital era. Digital information resources are important components of library information resources. After ten years development, a considerable amount of heterogeneous database based on different hardware and software have been formed in the digital library, which brings a lot of inconveniences to the users of digital resources. So, one of the most important thing that digital library need to face is to improve the interoperability among heterogeneous databases on the existing information infrastructure. Simultaneously, facing rapidly growing digital resource, it is very important for digital library to choose a suitable and reasonable architecture, which can support restore of information, search functions, function expansion and effective system maintenance better.With regard to the construction of the digital library which based on national defense industry, and on the basis of architecture design technology of digital library nowadays, this paper uses Struts framework based on MVC design pattern mode to design digital library, and has developed a distributed, scalable and multi-user documentation system.On the basis of researches on evolution of heterogeneous database, this paper brings forward a resource integration system which based on analyses of logical page. With a common service interface, system based on the basic principles of search engine realizes effective search of heterogeneous data in XML and web service distributed architecture integrally. System makes use of data acquisition, data optimization and integration of heterogeneous data to improve the accuracy and efficiency of integration, and eventually organizes all heterogeneous resources into one structure which can provide one-stop information search services for users. Compared with traditional methods of logical or physical integration, this method does not need to think about data source format, so function more simple and effective in time, space and complexity of algorithm.