Introduction to Software Test – Performance Test

Performance Testing

Performance Testing

Most software test activities are intended to measure the quality of a program, and if inadequate, document what needs to be fixed. There is another form of testing which really is not a ‘quality’ test at all.

That is ‘Performance Test’, and just like it sounds like, it is a measurement of performance, not quality. It is intended to show that the program is ‘fast enough’, not that it ‘works right’.

Stand out from the Crowd!

40% OFF – Mobile App QA e-Learning Course

Order Now

Because the output of this test is measurements, which will be used to claim a level of performance, or certify that a required level is met, the environment is key. It must be 1) exactly like the specified environment and 2) rigidly controlled. What good would performance measurements on a simulated system be? Or on systems or with/without other environmental aspects which do not match the target customer environments? And you want to be able to ensure that the measurements are ‘repeatable’.

The measurements can include times, either reported by an external timing device or by the system under test (the most accurate method). For instance, “Automatic Backup of a database containing 100,000 records at a time when the system is running at a load of 25% of normal, takes 14 minutes.” It can also include ‘throughput’, like “107 users can simultaneously submit requests without impact to the program” or “A user can submit as many as 19 requests in one minute”.

The output from this test is often used as sales material, or to prove that specifications were met. It can also be used to point out the need to rework the product for better performance, if the performance does NOT meet specifications, or does not make the product particularly competitive.

Thus there is some skill and some art in deciding what to measure. You need to measure that which will be important to the customer as well as practical to measure.