A parallel approach for improving Geo-SPARQL query performance

Abstract

Geospatial Semantic Web promises better retrieval geospatial information for Digital Earth systems by explicitly representing the semantics of data through ontologies. It also promotes sharing and reuse of geospatial data by encoding it in Semantic Web languages, such as RDF, to form geospatial knowledge base. For many applications, rapid retrieval of spatial data from the knowledge base is critical. However, spatial data retrieval using the standard Semantic Web query language -Geo-SPARQL -can be very inefficient because the data in the knowledge base are no longer indexed to support efficient spatial queries. While recent research has been devoted to improving query performance on general knowledge base, it is still challenging to support efficient query of the spatial data with complex topological relationships. This research introduces a query strategy to improve the query performance of geospatial knowledge base by creating spatial indexing on-the-fly to prune the search space for spatial queries and by parallelizing the spatial join computations within the queries. We focus on improving the performance of Geo-SPARQL queries on knowledge bases encoded in RDF. Our initial experiments show that the proposed strategy can greatly reduce the runtime costs of Geo-SPARQL query through on-the-fly spatial indexing and parallel execution.

Current and Future Challenges of Geospatial Semantic Web

“…The parallel approaches (Zhang et al 2013 andZhao et al 2014) can also be combined with other strategies to further improve the query performance of Geospatial Semantic Web queries. For example, with the growth and widespread use of smart phones, in the future many applications may use smart phones to query the needed spatial feature information from the Geospatial Semantic Web for updating and sharing the geospatial information.…”

“…The proposed parallel approaches (Zhang et al 2013 andZhao et al 2014) can also be combined with the caching strategy to further improve the efficiency of query performance over the Geospatial Semantic Web. The frequently queried results can be cached on the Geospatial Semantic Web servers to allow the systems to respond as quickly as possible.…”

“…Recently we also proposed a novel approach to optimize and parallelize spatial joins in Geo-SPARQL queries (Zhao et al 2014). The novel idea of the proposed parallel approach is to separate spatial and non-spatial components in a Geo-SPARQL query.…”

“…Ravindra et al 2011;Anyanwu 2013). However, except Zhao et al (2014) there is no study to deal with parallelizing spatial join computations to support efficient spatial RDF query, which is an important issue for the development of Geospatial Semantic Web (Zhang et al 2007).…”

Adaptive and Optimized RDF Query Interface for Distributed WFS Data

“…In the geospatial community, in order to facilitate discovery of geospatial web services compliant with OGC specifications and promote access to geospatial information via the OGC geospatial web services, efforts have been made to incorporate ontology into OGC geospatial web services (e.g., [15][16][17][18][19][20]). There are recent interests by researchers in exploiting the geospatial semantic web for automatic integration of semantically heterogeneous geospatial data (e.g., [19,[21][22][23][24][25][26][27][28][29][30][31][32]). …”