Reasoning with large ontologies stored in relational databases: The OntoMinD approach

A major obstacle to the development of ontologies in support of the Semantic Web is the poor capability of current ontology techniques to handle very large ontologies, in particular regarding scalability of reasoners. This paper builds on the assumption that very large ontologies can be efficiently handled using database management systems (DBMS), designed to provide best performance in storing, updating, and managing large volumes of data. To enhance DBMS with the reasoning functionality that characterizes ontology management, we propose to implement reasoning into the DBMS via a set of PL/SQL stored procedures. These procedures support all usual reasoning tasks: Class subsumption, property subsumption, class satisfiability. ABox consistency, and ABox realization. They perform these tasks at update time and materialize all inferred knowledge (facts and axioms) in the database. Contrarily to the inferencing at query time in most of existing works, our approach is designed to speed up ontology querying, which is supposed to represent the most frequent and therefore critical usage of ontologies. The paper discusses querying patterns and reports on benchmarking (with the LUBM benchmark) the performance of our prototype, called OntoMinD, compared to Oracle with Semantic Technologies. Benchmark results demonstrate the appropriateness of our approach. (C) 2010 Elsevier B.V. All rights reserved.

Published in:
Data & Knowledge Engineering, 69, 1158-1180

 Record created 2011-12-16, last modified 2018-03-17

Rate this document:

Rate this document:
(Not yet reviewed)