Materialized View-Based Processing of RDF Queries

The increasing interest in the RDF data model has turned the efficient processing of queries over RDF datasets to a challenging and crucial task. Indeed, the triple format of the RDF data model, along with the lack of structure that characterizes it, raise new challenges in data management both in terms of performance and scalability. In this paper, we consider improving the performance of RDF query evaluation by using materialized views. Starting from a workload of queries, we describe the space of possible views to materialize, we introduce ways for assessing the optimality of each view set and we propose practical algorithms for exploring the search space. When an RDF Schema is available, our algorithm takes advantage of it to guarantee the completeness of answering queries. We evaluate the efficiency of our search algorithms and demonstrate the benefits of the materialized views recommended by our algorithms on a fully implemented RDF querying platform.