Distributed and Parallel Path Query Processing for Semantic Sensor Networks

As the sensor networks are broadly used in diverse range of applications, Semantic Web technologies have been adopted as a means to manage the huge amount of heterogeneous sensor nodes and their observation data. Large amount of sensor data are annotated with spatial, temporal, and thematic semantic metadata. As a consequence, efficient query processing over large RDF graph is becoming more important in retrieving contextual information from semantic sensor data. In this paper we propose a novel path querying scheme which uses RDF schema information. By utilizing the class path expressions precalculated from RDF schema, the graph search space is significantly reduced. Compared with the conventional BFS algorithm, the proposed algorithm (bidirectional BFS combined with class path lookup approach) achieves performance improvement by 3 orders of magnitude. Additionally, we show that the proposed algorithm is efficiently parallelizable, and thus, the proposed algorithm returns graph search results within a reasonable response time on even much larger RDF graph.