Presented at: 8th International Semantic Web Conference (ISWC2009)http://data.semanticweb.org/pdfs/iswc/2009/paper290.pdf
RDF is an increasingly important paradigm for the representation of information on the Web. As RDF databases increase in size to approach tens of millions of triples, and as sophisticated graph matching queries expressible in languages like SPARQL become increasingly important, scalability becomes an issue. To date, there is no graph-based indexing method for RDF data where the index was designed in a way that makes it disk-resident. There is therefore a growing need for indexes that can operate efficiently when the index itself resides on disk. In this paper, we first propose the DOGMA index for fast subgraph matching on disk and then develop a basic algorithm to answer queries over this index. This algorithm is then significantly sped up via an optimized algorithm that uses efficient (but correct) pruning strategies when combined with two different extensions of the index. We have implemented a preliminary system and tested it against four existing RDF database systems developed by others. Our experiments show that our algorithm performs very well compared to these systems, with orders of magnitude improvements for complex graph queries.
DOGMA: A Disk-Oriented Graph Matching Algorithm for RDF Databases was presented at this event.
Keywords: Semantic Web
Resource URI on the dog food server: http://data.semanticweb.org/conference/iswc/2009/paper/research/290
Explore this resource elsewhere: