Castagna, Paolo; Dollin, Chris; Seaborne, Andy
Keyword(s): lubm, inference, rdf, parallel processing
Abstract: This report describes an implementation of a parallel inference engine for LUBM. Starting from the Univ- Bench ontology, we derive a specialized ruleset which captures its semantics. This ruleset can be simplified using simple transformations to remove data dependencies between rules. Finally, a parallel implementation of a custom LUBM reasoner, which performs inference in a streaming fashion using a small cluster, is described. Only preliminary results are presented, more experiments are necessary and an investigation on how to apply the same approach to OWL ter-Horst and OWL 2 RL profile is suggested.
External Posting Date: November 6, 2009 [Fulltext]. Approved for External Publication
Internal Posting Date: November 6, 2009 [Fulltext]