HDSW: Semantic sensor network system based on hadoop

Abstract

Ontology is used to annotate sensor network data and enhances its semantic, which is the motivation of Semantic Sensor Network (SSN). With the development of SSN, massive RDF datasets have been produced, so SSN is up against a big challenge now, i.e., the integration and application of data. In this paper, we present a Hadoop-based Semantic Sensor Network System (named HDSW) to achieve efficient data management for sensor data in RDF. The architecture of HDSW is composed of sensor network layer, persistent layer, Hadoop layer, function layer and UI layer. As one of the core component of HDSW, the RDF storage module is described in detail. The RDF storage module uploads RDF dataset to Hadoop by MapReduce, and stores BigRDF file by key/value mode. Then BigRDF is mapped to HBase table schema by using MapReduce. Finally, we evaluate HDSW system, and the experimental results show that it has good scalability and efficiency.