Safety First

You must be able to trust the test result.

When working with performance test you must be able to trust the result. The figures for answer times, throughput, or whatever numbers are key in your benchmarks and must be trustworthy. They affect customer confidence, SLAs, future requirements and can have a big economic impact if they are not correct.

Here are four key factors affecting a solution.

First factor is PureLoad. PureLoad has been used in customer solutions for performance test for more than 15 years and has proven stable and reliable from small load volumes to simulating millions of users.

The second factor lies in the solution engineering of the load test environment. When designing solutions, great care has to be taken so that the network itself (or network nodes) does not affect test results. There is a trade off between creating a realistic test environment for the System Under Test (SUT) and how much can be spent on the environment.

A third factor is the simulation of the traffic. The chosen traffic profile will have a large impact on the behavior of the SUT. For new applications the traffic profile may be unknown and assumptions have to be made but often there are similar applications from which data can be obtained. The chosen traffic model and how to work with changes to it is vital to getting trustworthy test results.

The creation of test cases is the fourth key factor. This is about knowing how your system is going to be used: what actions are mostly done, what is the intended use of the system, etc. This is normally detailed in user stories or protocol specifications depending on SUT.