Replaying execution trace models for dynamic modeling languages

Abstract

Back-end analysis tools aiming to carry out model-based verification and validation of dynamic behavioral models frequently produce sequences of simulation steps (called execution traces) as their output. In order to support back-annotation of such traces, we need to store and replay them within a modeling environment (outside the analysis tool). In the paper, we present a technique for replaying recorded execution traces of dynamic modeling languages. Our approach complements static and dynamic metamodels by introducing a generic execution trace metamodel which is used to replay completed executions of a simulation directly over the dynamic model. Furthermore, we present a technique to drive a simulation according to execution trace models. Our approach will be exemplified by the modeling language and trace information of the SAL model checker and BPEL business processes.