Rethinking performance testing

1. Performance testing why do it

Performance testing – Why do it?

Why do we performance test? The reason is that poor application performance can cause serious damage to a business. It is a risk. The sole purpose of performance testing is to mitigate that risk. Sometimes a tremendous risk, sometimes a small risk. But what do these risks look like? Let's have a closer look and distinguish three main areas: 

  1. The quality (performance) of the code
  2. The balance between capacity supply and demand
  3. Defects in the stack of hardware, software and parameters

Each of these areas can cause damage. Especially the first one can cause time consuming redesign.

This observation may look innocent, maybe even boring, but provides a paradigm shift that paves the way for new and more sophisticated testing methods and tools. Imagine an approach that splits into these three parts addressing each of the risks separately. The first part is to assess the time behaviour (response times) of the transactions. In the second part one determines how much capacity is required. The third part tests the stack of hardware, software and parameters for defects. Each part is done at the right time in the life cycle of an application, with adequate intensity and most of all at appropriate cost.