Faster query execution for partitioned RDF data

Abstract

This work demonstrates use of Materialized Views to enhance query performance for partitioned RDF data. Given a query, our system determines which views or combinations thereof can be used to answer it. Break- even analysis for the proposed system has been done based on view materialization and refreshment costs. The system performance was evaluated for 7 query types, 3 having Sub-Obj joins. It shows that our approach reduces query response time by an average of 26% for all query types w.r.t response time using just vertical partitioning. Specifically, for queries with Sub-Obj joins, the average reduction is by 37%. On scaling data up 8 times, the reduction changed from 37% to 79% for queries with Sub-Obj joins, and from 26% to 51% on an average for all query types. With the proposed technique, Semantic Web Applications shall be more interactive since queries having Sub-Obj. joins are expected for them. © 2013 Springer-Verlag Berlin Heidelberg.