Ulysses: an Intelligent client for replicated Triple Pattern Fragments

Thomas Minier (1, 2) , Hala Skaf-Molli (1, 2) , Pascal Molli (1, 2) , Maria-Esther Vidal (3, 4)

1 LS2N - équipe GDD - Gestion de Données Distribuées
2 LS2N - Laboratoire des Sciences du Numérique de Nantes
3 TIB Leibniz Information Centre For Science and Technology University Library
4 Fraunhofer IAIS - Fraunhofer Institute for Intelligent Analysis and Information Systems

Ulysses is an intelligent TPF client that takes advantage of replicated datasets to distribute the load of SPARQL query processing and provides fault-tolerance. By reducing the load on a TPF server, Ulysses improves the Linked Data availability and distributes the financial costs of queries execution among data providers. This demonstration presents the Ulysses web client and shows how users can run SPARQL queries in their browsers against TPF servers hosting replicated data. It also provides various visualizations that show in real-time how Ulysses performs the actual load distribution and adapts to network conditions during SPARQL query processing.

Mots clés: Load balancing; Data Replication; Intelligent client; Semantic Web; Triple Pattern Fragments; Fault tolerance

Domaines: Calcul parallèle, distribué et partagé [cs.DC]; Web

Thomas Minier, Hala Skaf-Molli, Pascal Molli, Maria-Esther Vidal. Ulysses: an Intelligent client for replicated Triple Pattern Fragments. The Semantic Web: ESWC 2018 Satellite Events (Awarded Best demonstration paper), Jun 2018, Heraklion, Greece. ⟨10.1007/978-3-319-98192-5_34⟩. ⟨hal-01789410⟩