Thoughts about using Constraint Solvers in Action

Abstract

SMT solvers power many automated security analysis tools today. Nevertheless, a smooth integration of SMT solvers into programs is still a challenge that lead to different approaches for doing it the right way. In this paper, we review the state of the art for interacting with constraint solvers. Based on the different ideas found in literature we deduce requirements for a constraint solving service simplifying the integration challenge. We identify that for some of those ideas, it is required to run large scale experiments for evaluating some of the ideas behind the requirements empirically. We show that the platform is capable of running such an experiment for the case of measuring the impacts of seeds on the solver runtime.