Updated: 20-October-2007
Creating a benchmark to provide a consistent, comparative measure of JMS-based messaging server performance is not simple. The challenges are numerous, for example:
To be useful and reliable, a JMS messaging benchmark must fulfill several
fundamental requirements:
And in the end, you hope you have something that is fair, relevant, and
understandable, that will address technology for a period of years.
Such is the case with SPECjms2007, which is the first industry-standard
benchmark designed specifically for enterprise messaging servers based
on JMS. It provides a consistent workload and performance metrics for
competitive product comparisons, as well as a framework for in-depth
performance analysis of enterprise JMS messaging.
All of this is due to the efforts of a team of people from around the world, that SPEC would like to thank (with apologies and thanks to those who may not be listed).
SPECjms2007 was developed by the SPECjms working group which is part of
SPEC’s Java subcommittee with the participation of Technische Universit�t
Darmstadt, IBM, Sun, Oracle, BEA, Sybase and Apache.
Thanks to all members of the SPECjms working group, in particular:
Samuel Kounev from TU Darmstadt for serving as project manager, benchmark
architect and working group chair,
Kai Sachs from TU Darmstadt for serving as benchmark architect, designer
and lead developer,
Marc Carter from IBM for serving as responsible developer for the design
and implementation of the driver, automation, and reporting framework,
George Tharakan from Sun Microsystems for his invaluable contributions
to the specification and design of the workload as well as to the definition
of the run and reporting rules,
Martin Ross from IBM, Binu John, Eileen Loh, Ken Dyer and Sagar Shirguppi
from Sun Microsystems, Russell Raymundo and Tom Barnes from BEA, Anoop
Gupta from Oracle, Sebastian Frischbier from TU Darmstadt and Evan Ireland
from Sybase for their hard work on testing, profiling and debugging the
benchmark,
Adrian Co from Apache for his excellent work on implementing the reporter,
Tim Dunn from IBM, Saraswathy Narayan from Oracle and Tom Barnes from
BEA for their contributions to the specification of the benchmark run
and reporting rules.
Many thanks also to
Alejandro Buchmann from TU Darmstadt, Lawrence Cullen, Robert Berry, Alan
Adamson, John Stecher and Matt Hogstrom from IBM, Steve Realmuto from
BEA and Ricardo Morin from Intel for their continued support of the SPECjms2007
project.
The German Research Foundation (Deutsche Forschungsgemeinschaft) for funding Samuel Kounev's work as part of grant No. KO 3445/2-1.
We would also like to thank:
Finally, we thank all of the people behind the scenes in architecture
groups, product development and performance groups who supported their
work for SPEC at their respective companies.
Samuel Kounev, Release Manager, SPEC Java Subcommittee