Menu

Introduction

Robert Vesse

Introduction

SPARQL Query Benchmarker started life purely as a SPARQL over HTTP benchmarking tool. Over time it has evolved into a general purpose API that while designed primarily for testing remote SPARQL servers (including queries, updates and GSP requests) can in principal be extended to test much more than that. It also nowadays includes support for In-Memory Testing that makes it possible to eliminate much of the communication overhead from your testing.

The API is described by javadoc available at http://sparql-query-bm.sourceforge.net/javadoc/ and in the documentation on this wiki.

Documentation

NB The documentation on this wiki refers to the heavily revised 2.x releases of the tool.

The following documentation topics are available:

  • Operation Mixes - Describes our notion of an operation mix, how to define a mix and the types of operations supported as well as how to define your own custom operations
  • Benchmarking Methodology - Describes our benchmarking methodology used by the benchmark command
  • General Notes - Describes general notes about some of the supported options and APIs
  • API - Describes how to use the programmatic API to run tests
  • In-Memory Testing - Describes how to perform in-memory testing both via the API and command line

Command Line

The command line currently comprises three main tools:

  • benchmark - Runs a benchmark of operations against a SPARQL server.
  • soak - Runs a soak test of operations against a SPARQL server.
  • smoke - Runs a single run of operations to see if anything fails
  • stress - Runs a set of operations repeatedly under progressively increasing load to see how a system reacts under high load
  • operations - Provides command line help with available operations

See the CLI documentation for examples of running the various commands.

Download

Sample Results

See some example results on the SP2B Results Page

Getting Support

For help and support please look at the documentation on this wiki, within the release or use the project mailing list sparql-query-bm-users@lists.sf.net

Developer Builds

You can obtain the artifacts as Maven SNAPSHOTs by following the instructions on the Maven page.

License

SPARQL Query Benchmarker is open source software developed by YarcData, a division of Cray Inc. It is licensed under a standard form of the BSD 2.0 License, see source for more details.

While Cray is making this software freely available and our intention is to foster community adoption of this tool we make no guarantees that we shall continue to maintain and develop this tool long term. Users may submit bug reports and patches via the mailing list but Cray will be under no obligation to accept these.

Acknowledgements

SPARQL Query Benchmarker relies on libraries and resources from several other open source projects:

  • Apache Jena - Apache Software License v2
  • SP2B - Queries only, BSD License
  • LUBM - Queries only, from research paper cited below

GUO, Y., PAN, Z., HEFLIN, J.. LUBM: A Benchmark for OWL Knowledge Base Systems. Web Semantics: Science, Services and Agents on the World Wide Web, North America, 3, mar. 2011. Available at: http://www.websemanticsjournal.org/index.php/ps/article/view/70/68. Date accessed: 01 Jun. 2012.


Related

Wiki: API
Wiki: BenchmarkingMethodology
Wiki: GeneralNotes
Wiki: In-Memory
Wiki: Maven
Wiki: OperationMix
Wiki: SP2B Results Page