Extracting runtime monitors from tests

Abstract

The effort and training required to specify runtime monitors in industry might be one of the factors explaining the slow uptake of the technique. In this position paper, we discuss ways in which runtime monitors can be extracted automatically from tests which are typically available with industrial systems. This brings with it a number of challenges, however, mainly because tests are typically focused on checking very specific behaviour, rendering the checks practically unusable in a runtime verification setting where the behaviour is user-directed rather than test-specified. In this context, we present an experiment and preliminary results which should provide the basis for further future exploration.