Bob Williams, Senior Manager e-Commerce Marketing at Harland Clarke, has a nice post about performance testing, load testing, and stress testing. He titles it Customer focused eCommerce: Volume testing techniques
I like his perspective and wish he would have added more about the impact of website downtime. It would be interesting to hear more about how a e-Commerce marketing executive calculates the cost implications. For example, the wasted marketing dollars when the site goes down under too many concurrent users, especially when caused by a splendid marketing campaign that is wildly successful in driving traffic.
It is also interesting to see yet another person differentiating between the 3 related testings: performance, stress, and load. He states, “The goal of load testing is not to break the system, but to observe how the system operates under a constant maximum load.” I agree that load testing is intended to break the system, however I would prefer to call the duration portion “endurance testing” or “soak testing”.
Finding memory leaks is tricky business, and my experience as a developer/tester would suggest there are benefits to separating load testing from endurance testing. They have different objectives. So, my recommendation is to delay attempting to find memory leaks until you have completed some load testing with acceptable response times and low error rates. Creating new test plans for endurance testing and allocating resources as a different sub-project has the benefits related to focus and specialization. It will just be better.
Well done Bob! Thank you for sharing your thoughts.