SemLAV: Local-as-View Mediation for SPARQL queries

Gabriela Montoya (1) , Luis Daniel Ibáñez (2) , Hala Skaf-Molli (2) , Pascal Molli (2) , Maria-Esther Vidal (3)

1 LINA - Laboratoire d'Informatique de Nantes Atlantique
2 GDD - Gestion de Données Distribuées [Nantes]
3 Departamento de Computación y Tecnología de la Información

Local-as-view mediators allow to query semantic heterogeneous data sources such as linked data endpoints, open data servers or any available web service. Nevertheless, the problem of rewriting conjunctive queries on LAV mediators is intractable in general, and the number of rewritings can be exponential on the size of the query. Most important, in the context of the Semantic Web even simple conjunctive SPARQL queries can be rewritten in millions of rewritings. In this paper, we propose SemLAV, a LAV mediation technique that relies on Semantic Web technologies to execute SPARQL queries against data sources. SemLAV implements a polynomial view selection algorithm to choose the relevant views that can answer a given SPARQL query. Additionally, SemLAV implements an execution strategy for gathering data from the relevant views and outputs a great portion of the answer in a short time. We present an experimental study for SemLAV, and compare its performance with execution strategies for LAV rewritings. The results suggest that SemLAV scales up to SPARQL queries and a large number of views, while it significantly outperforms existing solutions.

Mots clés: Data Integration; Local-as-view; Sparql Query

Domaines: Algorithme et structure de données [cs.DS]; Base de données [cs.DB]

Gabriela Montoya, Luis Daniel Ibáñez, Hala Skaf-Molli, Pascal Molli, Maria-Esther Vidal. SemLAV: Local-as-View Mediation for SPARQL queries. Transactions on Large-Scale Data- and Knowledge-Centered Systems, 2014, pp 33-58. ⟨10.1007/978-3-642-54426-2_2⟩. ⟨hal-00841985v1⟩