Integration of Semantically Aligned Heterogeneous Databases

Abstract

In this paper we propose a methodology whereby we semantically align heterogeneous databases without changing the databases. The alignment achieved is on the data instance level, which does not require the semantics of the data at the type level. We first introduce semantic alignment in the context of our databases research project then discuss in details how heterogeneous databases that are semantically aligned may be integrated. In general, we align heterogeneous databases with 'Global as View'. We show that by drawing on Barwise-Seligman's information flow channel theory and making use of Enhanced Entity-Relationship model, semantically aligned heterogeneous databases can be successfully integrated, which enables non-conventional queries on constituting databases. We show that without modifying databases our methodology integrates any chosen databases on data instances level without loss of information, which is our contribution over existing works. Furthermore, Barwise-Seligman's information flow channel theory represents a novel and mathematically sound approach to linking independent systems. As we use this theory as a theoretical basis of our work and therefore we believe that our ideas presented here have a general applicability in talking issues in distributed systems.