The ability of the RDF data model to link data from heterogeneous domains has led to an explosive growth of RDF data. So, evaluating SPARQL queries over large RDF data has been crucial for the semantic web community. However, due to the graph nature of RDF data, evaluating SPARQL queries in relational databases and common data-parallel systems needs a lot of joins and is inefficient. On the other hand, the enormity of datasets that are graph in nature such as social network data, has led the database community to develop graph-parallel processing systems to support iterative graph computations efficiently. In this work we take advantage of the graph representation of RDF data and exploit GraphX, a new graph processing system based on Spark. We propose a subgraph matching algorithm, compatible with the GraphX programming model to evaluate SPARQL queries. Some experiments are performed to show the system scalability to handle large datasets.