Scalable RDF triple store using summary of hashed information and Bit comparison

Abstract

In this paper, we proposed a scalable RDF triple store for massive-scale RDF data that processes the SPARQL query with many join operations in efficient manner. Graph characteristic of RDF data model hinders scalable and efficient indexing and querying over RDF triples. To address the problem, our query processing uses the pruning algorithm based on Bit-structure and summarized information to minimize data-reading. Our approach guarantees scalability and flexibility even for massive-scale RDF data by storing RDF triples in distributed fashion, providing the modifiable structure, and optimizing memory footprint of usage. The experiments shows that our system is better performing for queries with many join operations while uses less memory footprints.