Dependability benchmarking of computer systems aims to provide generic, reproducible methods for characterising the behaviour of a computing system in the presence of faults. A dependability benchmark must define: i) the target system and the benchmarking context, ii) the measures to be evaluated, as well as iii) the experimental conditions, ensuring that it is repeatable, reproducible, representative, portable, and cost-effective.

The STF group capitalises on the knowledge acquired during its participation in the Dependability Benchmarking project to define dependability benchmarks that are critical for decision-making when comparing alternative fault tolerance mechanisms or their configurations.