Using an index of precomputed joins in order to speed up SPARQL processing

SparQL is a query language developed by the W3C, the purpose of which is to query a data set in RDF representing a directed graph. Many free available or commercial products already support SparQL processing. Current index-based optimizations integrated in these products typically construct indices on the subject, predicate and object of an RDF triple, which is a single datum of the RDF data, in order to speed up the execution time of SparQL queries. In order to query the directed graph of RDF data, SparQL queries typically contain many joins over a set of triples. We propose to construct and use an index of precomputed joins, where we take advantage of the homogenous structure of RDF data. Furthermore, we present experimental results, which demonstrate the achievable speed-up factors for SparQL processing.