A development of RDF data transfer and query on Hadoop Framework

Abstract

A RDF graph is typically stored in XML file or relational database. However, when it becomes a large RDF graph, an alternative way to handle the storing and query RDF graph or linked data is to use MapReduce algorithm and Hadoop framework. In this paper, we propose a supporting tool to perform data transfer and query on big RDF graph. We intend to reduce the access time and query response time by using Hadoop Framework. The RDF/XML or linked data is converted into a huge set of N-triples and they are uploaded onto Hadoop storing in data nodes of Hadoop Distributed File System (HDFS). The query of RDF graph in terms of SPARQL is analyzed and converted into a specific N-triple format as to search the answer using Jena Algebra. The MapReduce algorithm is developed to relevantly manipulate the RDF graph.