A Benchmark for Conflict Detection Components of Model Versioning Systems

Abstract

In the past years, the model-based development paradigm enjoyed an ever increasing adoption in academia and industry projects. This increasing adoption lead to several diverse approaches of socalled model versioning systems that support collaborative model-based development: multiple developers apply concurrent changes to potentially overlapping parts of the same model and model versioning systems merge their changes to obtain a new consolidated version of the model. One of the most crucial components in such model versioning systems is responsible for detecting conflicts among concurrent changes to avoid the loss of changes and corrupted models. However, due to the diversity of existing conflict detection components, their relative quality in terms of functional and nonfunctional properties is hard to determine. To address this issue, we propose a benchmark enabling the automatic evaluation of their accuracy, as well as their execution time, whereas both state-based approaches and operation-based approaches are supported alike.