Scalability of Model Transformations: Position Paper and Benchmark Set

As model transformations are often considered the “heart and soul” of Model-Driven Engineering (MDE), the scalability of model transformations is vital for the scalability of MDE approaches as a whole. The existing research on scalable transformations has largely focused on performance behavior if the involved input models grow in size. In this work, we address a second key dimension for the practical scalability of model transformations: The effect as the transformation specification itself grows larger. We outline a number of challenges related to large model transformations, specifically affecting the quality concerns maintainability and performance. We introduce three model transformation benchmarks and discuss how they are affected by these challenges. The transformation rule sets in these benchmarks have acted as an evaluation basis in our previous work. Our objective is to establish a community benchmark set to compare model transformation approaches with respect to the aforementioned quality concerns.