A Constraint-Solving Approach for Achieving Minimal-Reset Transition Coverage of Smartcard Behaviour

Abstract

Smartcards are security critical devices requiring a high assurance verification approach. Although formal techniques can be used at design or even at development stages, such systems have to undergo a traditional hardware-in-the-loop testing phase. This phase is subject to two key requirements: achieving exhaustive transition coverage of the behavior of the system under test, and minimizing the testing time. In this context, testing time is highly bound to a specific hardware reset operation. Model-based testing is the adequate approach given the availability of a precise model of the system behavior and its ability to produce high quality coverage while optimizing some cost criterion. %l'argument n'est pas convainquant. This paper presents an original algorithm addressing this problem by reformulating it as an integer programming problem to make a graph Eulerian. The associated cost criterion captures both the number of resets and the total length of the test suite, as an auxiliary objective. The algorithm ensures transition coverage. An implementation of the algorithm was developed, benchmarked, and integrated into an industrial smartcard testing framework. A validation case study from this domain is also presented. The approach can of course be applied to any other domains with similar reset-related testing constraints.