Live linked data: synchronising semantic stores with commutative replicated data types

  • Authors:

  • Luis-Daniel Ibáñez;Hala Skaf-Molli;Pascal Molli;Olivier Corby

  • Affiliations:

  • Laboratory of Informatics of Nantes-Atlantique, University of Nantes, 2 rue de la Houssiniere, BP 92208, 44322 Nantes Cedex 3, France;Laboratory of Informatics of Nantes-Atlantique, University of Nantes, 2 rue de la Houssiniere, BP 92208, 44322 Nantes Cedex 3, France;Laboratory of Informatics of Nantes-Atlantique, University of Nantes, 2 rue de la Houssiniere, BP 92208, 44322 Nantes Cedex 3, France;INRIA Sophia Antipolis-Mediterranee, Polytech - Templiers 930, Route des Colles - BP145, 06903 Sophia Antipolis Cedex, France

  • Venue:
  • International Journal of Metadata, Semantics and Ontologies
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Linked Data is currently interconnecting information on the web, creating the web of data. It allows data consumers to combine different data sets and perform powerful queries. However, this means either to copy data sets locally or perform distributed querying. Local copies have problems of freshness, distributed queries, of scalability and performance. Linked Data producers are currently going live by providing streams of data updates, opening a third way to query: synchronise and search. Each Linked Data node can follow update streams of others, creating a social network of live updates: the Live Linked Data LLD. Unfortunately, synchronising data among autonomous participants raises issues of concurrency and consistency. In this paper, we propose SU-Set, a Commutative Replicated Data Type CRDT for RDF graph updated with SPARQL Update 1.1. We describe how a semantic store can use SU-Set to ensure eventual consistency in LLD, with a low overhead in time, space and communication.