A SPARQL Engine for Streaming RDF Data

The basic data format of the semantic Web is RDF. SPARQL, which has been developed by the W3C, is the upcoming standard for RDF query languages. Typical engines for processing SPARQL queries on RDF data first read all RDF data, may build indices of the complete read data and afterwards evaluate SPARQL queries. Such engines cannot operate on streaming RDF data. Streaming query engines operating on streams of data can (a) discard irrelevant input as early as possible, and thus save processing costs and space costs, (b) build indices only on those parts of the data, which are needed for the evaluation of the query, and (c) determine partial results of a query as early as possible, and thus evaluate queries more efficiently. We propose such a streaming SPARQL engine, which is the first streaming SPARQL engine to the best of our knowledge.