Performance Testing enables the testing of anticipated load and concurrency you would expect to encounter in the real world prior to release. Examples of this may be simulating thousands of hits on a site, checking it doesn’t crash and ensuring response time is acceptable. It may be non-concurrency related and actually involve checking a system can cope with a batch process submitting a large number of transactions. Performance testing really depends on the application under test – every application tends to differ a little. Its now common for performance testing to relate to simulation of concurrent users on the presentation tier of the client side of a website.
Why do performance testing?
In my experience the main reason can be quite simply summed up as this “To measure and report the behaviour of the of the website under anticipated live load” … by measuring the performance of a site we can report key items such:
- End user response times
- Key business processes
- CPU and Memory statistics
- Performance of a planned release compared to the current live system
The end output should give a high degree of confidence in the ability of the application to handle live volumes and patterns prior to go-live.
See also – Performance Testing and the different types