Research And Application Of Heterogeneous Database Integration Middleware Construction

With the rapid development of IT industry, the demand of sharing the data from existing systems is increasing, and there is an urgent need for an approach to integrate the distributed heterogeneous database resources. At present data warehouse and federated database are two heterogeneous databases integration solutions in common use. However, these solutions are not only complicated but also subject-oriented. Mean while, intermediary system solution is more flexible, low coupling, cheep, and database sources are still completely autonomous.The subject of this paper comes from an actual project - Dalian Municipal Communications Bureau data center building. Database heterogeneousness of mainly reflected in two aspects: one is different database management systems, there are different data access approaches; the other is different semantics, including database schema difference, data meaning and type conflicts. And the autonomy of these databases also impacts the integration. This paper studies the heterogeneous database integration middleware construction and its application in actual project. First of all, according Dalian Municipal Communications Bureau data characteristics, a strategy used for building global virtual database is described. Secondly new database resource should be registered. Web service based data access agent provides local database access service, and the schema mapping relation between local database and global virtual database is build to shield the semantic differences. Finally, on the basis of the above, a query mapping algorithm is described, it can transform a global virtual database query to a group of real local database queries, and merge the results from local databases.In this paper, heterogeneous database integration middleware uses web services to wrap the business, and take the XML format as data exchange standards. The middleware is applied in Dalian Municipal Communications Bureau data center building project, it can successfully integrate some heterogeneous databases, and build global virtual databases with uniform database access approaches.