Monday, 5/21/2007
5:00 PM - 6:00 PM
Level: Technical - Intermediate

AllegroGraph is a highly scalable and performant RDF store. Here we describe and demonstrate our implementation of SPARQL, discussing some of the characteristics of querying non-trivial datasets, and our approaches to optimization. Our SPARQL implementation is closely integrated with two new features in AllegroGraph 2.0: a built-in 'RDFS++' reasoner, and range queries over a wide variety of datatypes. SPARQL's filter operations can be cast in terms of these range queries, and query patterns implemented on top of heavily indexed cursors. Together, these provide both time- and space-efficient querying, even over very large datasets. This presentation will provide a demonstration and walkthrough of SPARQL on AllegroGraph, illustrating its performance and exploring its applications. It will also give a few behind-the-scenes glimpses into how queries are optimized for this kind of a system.

Richard Newman has extensive experience in researching, developing, and applying the Semantic Web stack, particularly SPARQL.