Monitoring the Execution of Declarative Model Transformations

Abstract

Model transformations, applied at design and run time, are key artifacts in Model-Driven Software Engineering. The monitoring of a transformation’s execution is a prerequisite to enable a software engineer to identify performance bottlenecks and improve transformations. Monitoring is particularly relevant for declarative model transformations since the order of execution is not explicitly defined but instead the result of internal heuristics of the transformation engine. In this paper, we present how we monitor the execution of Henshin model transformations using Kieker as well as the resulting monitoring overhead. We show that the monitoring overhead depends on the size of the input model and that it is between 17.03% and 28.44%.