RDFSpark: a new solution for querying massive RDF data using spark

The invasion of semantic data, the rapid growth of RDF data has brought significant new challenges in the querying of RDF data. On the other hand, Apache Spark is an open source distributed computing framework, characterized by its speed as MapReduce, Big Data processing has never been easier. In last years MapReduce solves problems at scales unimaginable a few years ago. But like any other tool, it remains limited. Several research works propose the querying of large volumes of RDF data using MapReduce as H2RDF, We find Spark which is more powerful and robust than MapReduce, and it is 100 times faster than MapReduce. In this paper, we present RDFSpark: a new distributed RDF query management based on Spark to ensure scalability, fault tolerance, and performance to solve low-efficiency problems for RDF data query. n this approach, a SPARQL query is analyzed first by the Parser, and compiled before being passed through a series of optimization techniques, then it will be translated to a Spark program. The results of experiments conducted on huge volumes of RDF data show that RDFSpark has a high querying performance.Â