Star Pattern Fragments: Accessing Knowledge Graphs through Star Patterns

• DOI: 10.48550/arxiv.2002.09172
• Publication Date: 2020-01-01
• Authors: Aebeloe, Christian; Keles, Ilkcan; Montoya, Gabriela; Hose, Katja
• Keywords: FOS: Computer and information sciences; Computer Science - Databases; 0202 electrical engineering, electronic engineering, information engineering; Databases (cs.DB); 02 engineering and technology

The Semantic Web offers access to a vast Web of interlinked information accessible via SPARQL endpoints. Such endpoints offer a well-defined interface to retrieve results for complex SPARQL queries. The computational load for processing such SPARQL endpoints offer access to a vast amount of interlinked information. While they offer a well-defined interface for efficiently retrieving results for complex SPARQL queries, complex query loads can easily overload or crash endpoints as all the computational load of answering the queries resides entirely with the server hosting the endpoint. Recently proposed interfaces, such as Triple Pattern Fragments, have therefore shifted some of the query processing load from the server to the client at the expense of increased network traffic in the case of non-selective triple patterns. This paper therefore proposes Star Pattern Fragments (SPF), an RDF interface enabling a better load balancing between server and client by decomposing SPARQL queries into star-shaped subqueries, evaluating them on the server side. Experiments using synthetic data (WatDiv), as well as real data (DBpedia), show that SPF does not only significantly reduce network traffic, it is also up to two orders of magnitude faster than the state-of-the-art interfaces under high query load.