Load Testing Manually

As I sit here on this Monday morning and try to figure out what I should do first, it hits me that today is the beginning of an exciting year. 2009 will probably be remembered by most people as the big global economic recession. For me, 2009 will be significant because of the launch of LoadStorm.

We are proud of releasing a product that has received such wonderful response. Our goal was to solve the load testing problem for web developers like us. We needed a tool that doesn’t require a big budget, and we didn’t have time/inclination to learn a new scripting language. The frustration to properly configure an unknown tool was not on my to do list. Thus our team said, “Let’s fix that.”

This morning I did a quick check and I’m happy to report that 3,736 people can now be called LoadStorm customers. We are looking forward to a great year because we expect to continue helping the most developers with the least expensive load testing solution.

Load Testing Manually

David Makogon provided us with an interview in 2009, and he has been conducting web load testing for over a decade. This morning I opened an email from David that contained an amusing story that I would like to share:

Back in the day, a client brought us in for load-testing. They had some arbitrary goals (like 100 concurrent users, 5-second response max, etc.). They informed me they already tested with 20 concurrent users, and everything worked fine. When asked how they tested 20 concurrent, they explained that they set up 20 computers in a lab, had one person at each computer, with browser open. “On the count of three, everyone pushed Enter at the same time. No issues.”

Needless to say, once we fired up the “real” load-testing tool, we broke the site with “two” concurrent users, due to session management code that wasn’t thread-safe….

Although I hate to admit it, we at CustomerCentrix have load tested like that too. It was probably 1999 when we were developing a web conferencing tool for LearnCentrix synchronous classes. We called about 25 of our closest friends to see if they would participate in a web classroom with us. The Java applet-based conferencing didn’t match our performance testing objectives. After that, getting 25 people to help us load test just wasn’t feasible. Oh the fun of the early web days….

Thanks to David for sharing. Please send us any stories you have about load testing or performance testing. Or if you are interested in participating in an email interview, please email me at [email protected].