Mobile Agent And Xml-based Virtual Database

Virtual database technology is a new arisen technology of data integration in the last few years, it is the inter-application result of computer network and database technology. The aim of constructing virtual database is to provide a united relational view for user transparently visiting Internet sites and other data sources. Virtual database view Internet and other data sources as a single database, and make it becomes a part of enterprise or Internet infrastructure.In view of some problems existed in the application of virtual database, this paper proposes an improvement scheme-virtual database framework based on mobile agent and XML. Making use of powerful function of mobile agent and XML to virtual database, not only can reduce the amount of data transmission in Internet and response time of integration system, allow network' s disconnection in some case, but also can simplify the implementation of integration system, speed up spread and share of information resources, hasten the innovation and development of virtual database technology. This paper discusses some key technologies of virtual database framework in detail.Query processing is a core problem of virtual database management system (mediator). This paper uses query processing as a main line, discusses and researches query rewriting , query optimization technology. As virtual database uses relational data model, the paper makes a special effort to discuss query rewriting technology -rewriting queries using views and its conjunctive query, containment mapping under this model. Query optimization is the main problem of query processing research, thus in particular discusses cost estimated, query plan, search spaces and so on related problems.Wrapper is an important part of virtual database system. This paper proposes and researches a wrapper generation method based on sample example. By observing related Web data sources, simulating user's navigation behaviors, studying how to fetch result pages in demand under the guidance of heuristic rules. The whole learning process includes identifying search form, filling in and submitting search form, fetching result pages from return pages. After getting result pages, using an algorithm-object extraction algorithm, which can extract data in demand from the result page. This algorithm views data in page as a complex object that contains attributes and has a hierarchical structure. Firstly extracting related attributes, then assembling these attributes as a data object, thus completing extraction of data in demand. As experimental results show, this method achieves expected goal in effectiveness and expressiveness.