Sangam: A Framework for Modeling Heterogeneous Database Transformations

A broad spectrum of data is available on-line in distinct heterogeneous sources, and stored under different formats. As the number of systems that utilize the heterogeneous data sources grows, the importance of data translation and conversion mechanisms increases greatly. The goal of our work is a to design a framework that simplifies the task of translation specification and execution. Translation specification between the source and the target schema can be accomplished via (1) the discovery of matches between the source and the target schemata; (2) the application of a pre-defined translation templates; or (3) via manual user specification. In this paper we present a flexible, extensible and re-usable translation modeling framework wherein users can (1) explicitly model the translations between schemas; (2) compose translations from an existing library of modeled translation patterns; (3) choose from a library of translation operators; (4) generate translation models based on a match process; (5) edit such translation models; and (5) for all of these translation models, choose automated execution strategies that transform the source schema and data to the desired target schema and data. In this paper, we present the system architecture for such a translation modeling framework.