KeY + LARVA = STARVOORS

Abstract

Over the past decades, various forms of automated verification techniques have been proposed and explored in the literature, mostly falling in one of two categories — static and dynamic verification. On one hand, static verification techniques have the ability to verify properties across all possible executions of a system, but fully automated verification is typically not easy to perform. On the other hand, dynamic verification techniques, such as runtime verification, can only give feedback on single executions of the system, but are easy to automate. However, despite the fact that monitoring is typically easy to automate, its adoption in industry is limited — primarily due to the overheads in the system execution that such monitoring typically induces. In this paper, we explore a novel approach in which we combine the two approaches — using static analysis to prune parts of the specification, thus reducing the overheads for the dynamic verification process. We present our framework STARVOORS, which embodies this approach by combining the static analysis tool KeY and the dynamic verification tool LARVA, and discuss how it has been applied to the Mondex case study — an electronic purse implementation. The results presented here have been accepted for publication in and builds upon our previous results from.