Triple stores are the backbone of the evolution of the Linked Data Web. However, even standard queries on many attributes can result in high query response times using available triple stores. For use in commercial search applications, particularly for web scale applications, short and predictable query response times are a strict necessity in order to achieve industrial applicability. In this paper, a custom index extension for the Virtuoso triple store is used to improve the runtime of typical queries. The index extension is connected to a lightweight bitset index that covers the properties in question. A case study with real world data in the e-commerce domain shows that more comprehensive queries even on highly materialized data result in consistently improved query response times with a predictable upper bound that meets industrial requirements. Hence, to the best of our knowledge, industrial applicability is achieved for the first time considering the requirements of the e-commerce industry.
[1]
Abraham Bernstein,et al.
Hexastore: sextuple indexing for semantic web data management
,
2008,
Proc. VLDB Endow..
[2]
James A. Hendler,et al.
Matrix "Bit" loaded: a scalable lightweight join query processor for RDF data
,
2010,
WWW '10.
[3]
Gerhard Weikum,et al.
Scalable join processing on very large RDF graphs
,
2009,
SIGMOD Conference.
[4]
Nieves R. Brisaboa,et al.
Compressed k2-Triples for Full-In-Memory RDF Engines
,
2011,
AMCIS.
[5]
Jens Lehmann,et al.
DBpedia SPARQL Benchmark - Performance Assessment with Real Queries on Real Data
,
2011,
SEMWEB.
[6]
Orri Erling,et al.
RDF Support in the Virtuoso DBMS
,
2007,
CSSW.
[7]
Gerhard Weikum,et al.
The RDF-3X engine for scalable management of RDF data
,
2010,
The VLDB Journal.