Question regarding Load/Stress Testing with Jiffy

18 views
Skip to first unread message

perform67

unread,
Aug 15, 2008, 3:02:16 PM8/15/08
to jiffy-web
I've been doing some performance testing of an AJAX application using
Jiffy and, so far, have been very pleased with the results it
provides. I'm finding that it paints a much more accurate picture of
"real world" system performance than tools that don't execute client
side javascript during test runtime (for example, SilkPerformer,
JMeter, Visual Studio Web tests). One advantage those tools appear to
have though is the ability to simulate a large number of users through
load or stress tests.

The question I have is if anyone is aware of a tool that simulates a
large number of users that actually execute client side javascript
code at runtime (i.e., simulate the activity of a browser)? My goal
being to use a tool to drive a 500+ users test on our system and have
jiffy log the marks and measures while that test is executing. The
only tools I'm aware of that actually execute the client side
javascript code during a test are geared towards single user
functional testing (for example, Selenium, SilkTest). It may be a
shot in the dark, but I'm hoping someone may know of a tool or means
of doing this that I'm not aware of.

Thanks in advance. Thank you also to those that have provided the
Jiffy framework/toolset -- it's been of great value to us thus far.

Devin Ben-Hur

unread,
Aug 15, 2008, 6:30:21 PM8/15/08
to jiff...@googlegroups.com
perform67 wrote:
> The question I have is if anyone is aware of a tool that simulates a
> large number of users that actually execute client side javascript
> code at runtime (i.e., simulate the activity of a browser)? My goal
> being to use a tool to drive a 500+ users test on our system and have
> jiffy log the marks and measures while that test is executing.

I haven't used it for stress testing, but WATiR
<http://wtr.rubyforge.org/index.html> might be capable of doing what you
want. You'll have to build your own framework to drive load tests over
its primitives, I think.

WATiR is a ruby library which drives MSIE and allows you to build tests
which control the browser and verify the resulting DOM. WATiR includes
a primitive to create multiple IE windows in separate processes, so I
imagine you could write some recipes and play them concurrently through
multiple windows to get your X users simulation.
<http://wiki.openqa.org/display/WTR/FAQ#FAQ-HowdoIopenmultipleIEwindowsinseparateprocesses%3F>

(Yes, I know IE is a limitation for posix developers; there's work in
progress porting WATiR to support FF and Safari
<http://wtr.rubyforge.org/platforms.html>.)

> Thanks in advance. Thank you also to those that have provided the
> Jiffy framework/toolset -- it's been of great value to us thus far.

It's gratifying to see people using it, and contributing back.

-Devin

Devin Ben-Hur

unread,
Aug 15, 2008, 7:24:18 PM8/15/08
to jiff...@googlegroups.com


On Fri, Aug 15, 2008 at 3:30 PM, Devin Ben-Hur <dbe...@gmail.com> wrote:
I haven't used it for stress testing, but WATiR <http://wtr.rubyforge.org/index.html> might be capable of doing what you want. You'll have to build your own framework to drive load tests over its primitives, I think.

I little more poking around reveals a Load-testing example in the WATiR wiki <http://wiki.openqa.org/display/WTR/Re-Usable+Load+Testing+Example>

Patrick McFadin

unread,
Aug 16, 2008, 11:28:31 AM8/16/08
to jiff...@googlegroups.com
We use a tool called Bad Boy. http://www.badboy.com.au/

It will record your clicks as you walk through a site and then play it back so it will cover pages with AJAX. We've used it for some load testing by using it's threaded test feature. However, to get any real load on the web server, you will need to use multiple test machines. Once you get past 10-20 threads on a single machine, you may see your test flat line because your poor client is maxed out. That's when you bring in their other tool called Wave Test Manager. It will allow you to setup multiple test machines and have them controlled from a single spot.

We've had loads of machines running the same test and it definitely hammered our web site.

Patrick
Reply all
Reply to author
Forward
0 new messages