Accessing data through ontologies with ONDA

Accessing distributed and heterogeneous data sources through domain-oriented conceptual models (ontologies) facilitates the development of smarter information systems, as per the Semantic Web vision. To apply this vision to industrial grade applications, concept-based query answering must be efficient and scalable. This paper describes ONDA , a technology developed at the IBM Center for Advanced Studies of Rome which enables intelligent access to legacy relational information systems by means of ontologies. With ONDA , users can map rich conceptual models to existing or new relational data sources, and can query actual data through business concepts instead of the specific schemas of the data sources. By adopting a language of limited expressiveness ONDA ensures sound and complete query answering to conjunctive queries with the same efficiency and scalability of the underlying relational sources. This paper illustrates ONDA 's basic principles, outlines the system architecture and implementation, and presents prototypical applications. Finally, further research and development plans are discussed.