Hello Mark!
I've used The Grinder http://grinder.sourceforge.net/ , mainly because
it seemed to have less "overhead" than JMeter for testing. I noticed
that quite a bit of the effort needed on a benchmarking tool is related
to distributing and initialising the test clients. You might want to try
this part on existing tools before writing your own, unless you're
content on having a single machine testing tool.
Additionally, we already used MINA in our grinder tests by distributing
the MINA jar with the test files.
-Eero
Mark wrote:
I am sure there are other tools that perform load testing/benchmarking, but
they are not based on MINA :)
My intent is to tightly couple MINA and use it as the foundation for the
testing. The idea is that if people have a system they want to test, they
can either use an existing handler or write their own and let the
acceptors/connectors do all the work. So leverage this testing tool off of
MINA's flexibility. I was just curious as to the pitfalls that I might run
in to or if anyone has thought about writing such a system before.
As always, thanks Trustin !!
--Mark
On Dec 13, 2007 6:44 AM, Trustin Lee <[EMAIL PROTECTED]> wrote:
Hi Mark,
On Dec 13, 2007 1:54 PM, Mark <[EMAIL PROTECTED]> wrote:
I have been toying with the idea of writing a test tool based on MINA.
This
tool will be similar to what jmeter offers. I have used jmeter for
numerous
programs over the years and was never really happy with how it worked if
you
need to test a protocol that jmeter does not support 'out of the box'.
I am
curious as to what the MINA community thinks about such a program.
Do you mean a benchmark software that uses a MINA codec? It sounds
like a good idea, but we will also need to think about distributed
load generation and many more advanced features that existing testing
frameworks provide. Considering that, I'd personally prefer to
provide some kind of plugin, extension or module for a known
performance testing framework.
I know JMeter doesn't satisfy your needs but there might be some other
performance testing framework that will fit best (even if it's not
ideal best).
However, on the other hand, if you have some good idea about
performance testing framework which tightly integrates MINA, I'd like
to see you go ahead and how it will look like. Please feel free to do
what you want to do. It will be an exciting project anyway, because
we might be able to provide some features of the advanced testing
framework while improving MINA to support you to implement such
features.
Cheers,
Trustin
--
what we call human nature is actually human habit
--
http://gleamynode.net/
--
PGP Key ID: 0x0255ECA6
