DBOWL: Towards a Scalable and Persistent OWL Reasoner

With the increasing development of real applications using Semantic Web Technologies, it is necessary to provide scalable and efficient ontology querying and reasoning systems. In this paper we present DBOWL, a persistent and scalable OWL reasoner. Ontologies are stored in a relational database, using a description logic reasoner to pre-compute the class and property hierarchies and to obtain all the ontology information (i.e. properties domain and range) which is also stored in the database. Furthermore, a simple but expressive query language has been implemented, which allows us to query and reason on these ontologies. In order to show the use of the tool, we present an example that generates a relational database to store the Univ-Bench ontology. Finally, in order to asses the performance of our tool, we use LUMB, a well known benchmark to compare repositories in the Semantic Web extending it with more complex queries, and we compare the results obtained with Minerva. The results show the efficiency and scalability of DBOWL and the power of the query language.