DMCA
Diversified Stress Testing of RDF Data Management Systems
| Citations: | 11 - 1 self |
BibTeX
@MISC{Aluç_diversifiedstress,
author = {Günes Aluç and Olaf Hartig and M. Tamer Özsu and Khuzaima Daudjee},
title = {Diversified Stress Testing of RDF Data Management Systems},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Abstract. The Resource Description Framework (RDF) is a standard for conceptually describing data on the Web, and SPARQL is the query language for RDF. As RDF data continue to be published across heterogeneous domains and integrated at Web-scale such as in the Linked Open Data (LOD) cloud, RDF data management systems are being exposed to queries that are far more diverse and workloads that are far more varied. The first contribution of our work is an indepth experimental analysis that shows existing SPARQL benchmarks are not suitable for testing systems for diverse queries and varied workloads. To address these shortcomings, our second contribution is the Waterloo SPARQL Diversity Test Suite (WatDiv) that provides stress testing tools for RDF data management systems. Using WatDiv, we have been able to reveal issues with existing systems that went unnoticed in evaluations using earlier benchmarks. Specifically, our experiments with five popular RDF data management systems show that they cannot deliver good performance uniformly across workloads. For some queries, there can be as much as five orders of magnitude difference between the query execution time of the fastest and the slowest system while the fastest system on one query may unexpectedly time out on another query. By performing a detailed analysis, we pinpoint these problems to specific types of queries and workloads.
Keyphrases
rdf data management system diversified stress testing rdf data good performance stress testing tool waterloo sparql diversity test suite query execution time first contribution popular rdf data management system query language varied workload sparql benchmark linked open data detailed analysis resource description framework indepth experimental analysis heterogeneous domain second contribution diverse query specific type magnitude difference
