Regresar

OneQL: An ontology-based architecture to efficiently query resources on the semantic web

Abstract:

The widespread explosion of Web accessible resources has led to new problems on the traditional tasks of query evaluation and efficient data access. With this is mind, we have developed the ontology-based OneQL system which provides optimization and query evaluation techniques to scale up to large RDF/RDFS documents and complex queries, i.e., queries of any shape and with a large number of triple patterns. Efficiency of OneQL relies on the following components: - Query optimization and evaluation techniques that focus on cost models to estimate the execution time of a plan, and on searching the space of plans of any shape, i.e., bushy plans can be generated according to their estimated cost. - Bhyper: A directed hypergraph-based representation of RDF documents to directly access triples that share the same subject and property values, or the same property and object values. We report on the quality of the developed strategies, and have observed that implementing RDF documents with Bhyper and producing low estimated cost bushy plans, can speed up the evaluation time by up to four orders of magnitude.