From time to time I see questions on different forums asking what skills are necessary for performance testers. There were pretty interesting discussions. Looks like most experts agree that performance testing requires more skills than just knowledge about how to create a script for a particular load testing tool. While it is still possible to imagine a performance tester in a large corporation with deep specialization who only creates scripts and mechanically runs them while other performance experts monitor the system and analyze results, I don't see many perspectives neither for this person, nor for the approach. Systems become so complicated now that the sum of specialized expert views doesn't give the whole performance picture.

Late last week, the Sydney Futures Exchange gave 45 minutes notice of a lunchtime outage for an "urgent hardware rectification". In a market reporting good financial results, how could this have any impact? According to one press article, the outage was because new hardware was installed. Another article reported a backup computer database was faulty, and replacement was a safer option than risking continued trading. I want to know who did this risk assessment. It was the heaviest trading day on record, and volumes have been increasing steadily, growing 20% in the last 12 months alone. The system itself is 18 years old, but hopefully the hardware is newer!

Theme: Critical Incidents in Effective Performance Testing

Conference Location and Time
MITRE Bedford MA, USA

April 12 – 14, 2008

Major Questions of Interest
Questions we are interested in exploring at WOPR8 include:

What critical Incidents lead to effective performance testing? By sharing and attempting to understanding the critical incidents leading to effective performance testing we will explore how our craft can make an important contribution to all IT and Development projects.

Attended CMG 2006. It was really good ! I am too lazy to write down everything that was there, but there are others: here is one detailed blog. Of course, it is just one view: there was up to 12 different tracks at each timeslot. Unfortunately papers are available to members only, but Measure IT newletter is free (requires free registration).

Another great article by Scott Barber Diagnosing Symptoms of Poor System Performance (see pp.14-20) continues to discuss practical approaches to diagnose problems. Together with the previous Scott’s article about the subject How to Identify the Usual Performance Suspects (see pp.20-29) it is a must reading for everyone testing performance.

I forgot that many corporate spam filters don't allow EXE downloads.Here are two alternatives to get the stress testing movie: ZIP file (1.1 meg)Flash presentation (7 meg)

"Instant stress testing" is an idea I frequently use when I am trying to find bugs fast.Download and run this EXE to see the video demonstration of this little test technique.

Mike Kelly builds on Scott Barber's work to show how you can combine performance-degradation curves and complex performance scenarios to help determine "good enough" quality for an application in terms of performance. Read the article here.