Implementation Of Multi-Datasources Wrapper On Data Integration Platform

With the development of informatization in modern society, enterprise activities usually need to extract data from multitudes of heterogeneous, distributed and autonomous data sources. It has become an important research issue in modern enterprise data integration field of how to mask the heterogeneous datasources, map the heterogeneous datasources into a uniform data model, and provide high-level application with access to data efficiently and effectively.Based on the idea of middleware model, this paper includes the research of relevant technology on multi-datasouces mapping in data integration platform. Based on Eclipse platform, aiming to develop multi-datasources integration framework, this paper has designed and realized a multi-datasources wrapper. The wrapper is composed of a datasource configuration module and a query module, which are used for configuring and querying multi-datasources. The datasource configuration module is developed as an Eclipse plug-in, which is used for configuring datasources by users in the form of a wizard, getting the metadata of datasources, submitting the metadata to the data service development platform, for which to develop data services, and at the same time generating a datasource configuration file automatically. The query module provides a common interface to data service management module. It is able to analyze the datasource configuration file and query script file, get the query and datasources information, transform XQuery to SQL equivalently by a XQuery analyzer, map the query request to the datasources and transform the query result to the XML format. Finally, the extensibility of the wrapper is discussed and the function of the wrapper illustrated by a specific example.To sum up, this paper has conductd a series of work from the designing of data integration framework and multi-datasources wrapper. An efficient, strong-expansible wrapper which covers multi datasources including databases, XML files and Text has been realised. It provides a good tool for heterogeneous system integration. Finally the paper discusses the future work of improving the integration framework.