Abstract
Proliferation of RDF data on the Web creates a need for systems that are not only capable of querying them, but also capable of scaling efficiently with the growing size of the data. Parallelization is one of the ways of achieving this goal.
There is also room for optimization in RDF processing to reduce the gap between RDF and relational data processing. SPARQL is a popular RDF query language; however current engines do not fully benefit from parallelization potential.
We present a solution that makes use of the Bobox platform, which was designed to support development of data-intensive parallel computations as a powerful tool for querying RDF data stores. A key part of the solution is a SPARQL complier and execution plan optimizer, which were tailored specifically to work with the Bobox parallel framework.
The experiments described in this paper show that such a parallel approach to RDF data processing has a potential to provide better performance than current serial engines.