Semantic information processing system based on CAN

Abstract

Distributed RDF repositories have become a hot spot of Semantic Web, distributed hash table (DHT) mecha-nisms provide efficient search capabilities without storing global indexing and are hence scalable to large RDF repositories. Existing systems mainly focus on distributing triples uniformly, however, geographic data has its particularity that there are many topological relations between spatial objects, which causes existing systems are not well-suited for geographic data. In this paper, we introduce a scalable Peer-to-Peer semantic information processing system based on content addressable network (CAN), the system focuses on spatial data and preserves spatial locality information by storing triples connected with its coordinates in the real world. Our system supports binary topological relations matching and topological range queries. The experiments demon-strates that the system is scalable and performs better than state-of-the-art systems in the field of geospatial semantic.