Towards Parallel Processing of RDF Queries in DHTs

Efficient RDF query evaluation is a requirement if the vision of a Semantic Web is to come true one day. Peer-to-peer (P2P) networks, specifically distributed hash tables (DHTs), could be one of the enabling technologies for the Semantic Web. Several DHT based RDF triple stores have been proposed but despite their differences they all process queries sequentially and hence leave an untapped potential for parallelism. This paper proposes and analyzes a scheme for parallel processing of RDF queries in DHTs. We describe a planning algorithm that attempts to reduce the response time of a query by considering parallel query plans prior to query execution. We provide an experimental evaluation of our approach and compare its results to an already existing scheme, i.e. the one of the BabelPeers project.