Comparaison théorique et empirique de systèmes de bases de données sémantiques

Ontologies are more and more used in information systems. When the size of data described by ontologies can not be manage in main memory, information system designers have to use specialized databases called semantic databases. As these databases use different architectures, storage layouts and ontology languages, this task is complex. In this paper, we propose a comparison and evaluation of these systems in order to help information system designers identify a relevant semantic database. We first present a state of the art on ontologies, their languages and storage layouts and architectures of associated databases. Then we propose a model for semantic databases that capture their diversity. Finally we provide a cost model and performance evaluation of six industrial and academic semantic databases. MOTS-CLÉS : Ontologie, Base de données, Modèle de coût, Evaluation de performances