A hybrid approach to storage scheme for semantic Web metadata

In the near future, it is anticipated that enormous semantic Web metadata described in RDF will be appearing. Thus, some researches on the management for the semantic Web metadata have been proposed. In most previous approaches, the semantic Web metadata as a form of triple is stored in a large relational table. Since it always requires the whole table to be scanned for processing a query, it may degrade retrieval performance. In this paper, we distinguish some frequently appeared properties in the semantic Web metadata and/or frequently used properties in a user query. RDF data with these distinguished properties are independently treated and stored in corresponding property-based tables respectively. For processing a query having a specific property, we can avoid full scanning of the whole data and only have to access a corresponding table. Finally, we partially implement and evaluate the proposed scheme. The initial results show that the proposed scheme achieves better performance for queries having a specific property.