Detecting Feature Interactions in Telecommunication Services by Bounded Model Checking

Feature interaction refers to situations where a combination of different services behaves differently than expected from the single services’ behaviors. In today’s intelligent telecommunication networks, feature interaction is considered a major obstacle to the introduction of new features. We propose a new interaction detection approach which uses bounded model checking. The central idea behind bounded model checking is to reduce the model checking problem to the propositional satisfiability (SAT) checking problem and to look for counterexamples that are shorter than some fixed length k for a given property. In the literature, it has been reported that this method can work efficiently; but the original method does not work well for feature interaction detection, because a large formula is required to represent the transition relation, thus resulting in large execution time. We develop an alternative encoding method to address the problem.