| || |
I have been tasked with implementing Microsoft’s Web Application Stress tool (formally Homer) for QA on projects developed by my firm. I would like some input from any of you out there that have used this tool. I am also looking for some insight in how you have quantified your test results for your customers or firm.
The problem I see with load testing is that it is somewhat of a ‘Black Art’. A customer will generally come to you and say I want to be able to support ‘x’ number of users, yet they do not really know what this means. I can easily use WAS to make a technical call on when a site is deteriorating or in danger of failure, but it is hard to take those numbers and convert them back into actual users.
I would appreciate any comment or insights.
Re: MS WAS/Homer
I always ask a vp of development - how many users do you antcipate being on your website or application during normal hours and at the peak of operations?
That is usually a good starting place.
After you have a number, you can run the performance tests, then continue to scale up until the app crashes or you reach your vu limit.
When the app goes into production, real time traffic numbers should match the testing tool results (I usually benchmark 5 or 10 vu to start out with).
Does this help?