Research And Implementation Of Data Distribution Technology For Information Integration Applications

In the current information integration applications, on the distribution of the updated data of the data sources is only for a single table generally, which reduces the transparency and security, for that, we consider a partial encapsulation for the data source at the middle layer of the data integration. By making use of the views that defined on the basic tables, on one hand, users can customize data; on the other hand, the views can hide the details of the data. In this paper, we analyze how to distribute the incremental data to multiple destinations efficiently at the heterogeneous and multi-source environment, specially focus on the computation of distribution data and management of sharing resources.In the implementation of system, we apply publish/subscribe model, which defines the data source as the publisher, the publication views as exporting relation and the destination database or application that require integrating the exporting relations as subscriber. The subscribers apply for a subscription from a middleware since the middleware separates the publishers from the subscribers, which decouple the two. In addition, the original relationship between the data source and the destination database from many-to-one expands to many-to-many because of using publish/subscribe model. Note that the system offers three ways (period, schedule and requirement) to distribute data to the subscribers.