Robust performance testing for digital forensic tools

In previous work, the authors presented a theoretical lower bound on the required number of testing runs for performance testing of digital forensic tools. However, experimental errors are inevitable in laboratory settings, occurring as measurement errors or as random errors and can result in practical situations where the number of testing runs is far from the theoretical bound. This paper adapts our former work to tolerate such errors in the testing results. The contribution of our new methodology enables the tester to achieve performance testing results of high quality from a manageable number of observations and in a dynamic but controllable way. This is of particular interest to forensic testers who do not have access to sophisticated equipment and who can allocate only a small amount of time to testing.