LHD: Optimising Linked Data Query Processing Using Parallelisation
Xin Wang, Thanassis Tiropanis, Hugh Davis · ePrints Soton (University of Southampton) · 2013
In the past few years as large volume of Linked Data is being publishing, processing distributed SPARQL queries over the Linked Data cloud is becoming increasingly challenging. The high data traffic cost and response time significantly affect the performance of distributed SPARQL queries as the number of SPARQL end point and the volume of data at each endpoint increase. In this context, parallelisation is promising to fully exploit the potential of connections to SPARQL endpoints and thus improve the efficiency of querying Linked Data. We propose LHD, a distributed SPARQL engine that is built on a highly parallel infrastructure and able to minimise query response time, and we evaluate its performance using a BSBM based environment.