Forschungsthemen
[MA] On the Energy-efficiency of Model Transformations
Within the Collaborative Research Center 912 – Highly Adaptive Energy-efficiency Computing (HAEC) – the Software Technology Group is investigating approaches to develop energy-efficient software. For this, an approach based on Model-driven Software Engineering (MDSE) has been developed. The approach allows to develop self-adaptive software, which automatically adapts itself to always run in its most energy-efficient configuration. While several studies on the energy-efficiency of various algorithms have already been conducted, studies on the energy consumption behavior of model-driven techniques are lacking.
Model Transformations (e.g., with the Atlas Transformation Language, ATL) are a key technique in model-driven software engineering. By now, model transformations are mostly applied at design time and, hence, their performance and energy-efficiency was not subject to improvements. But, with approaches as developed in HAEC, model transformations are conducted at runtime. Thus, approaches to improve the energy-efficiency of model transformations are required.
The following tasks have to be fulfilled:
- Compose an overview on concurrent model transformation technologies (ATL, QVT, etc.)
- Find and compare to related work
- Find literature on energy efficiency studies in the context of MDSE
- Find literature on performance studies of model transformation technologies
- Select at least 5 case studies from the ATL Zoo (http://www.eclipse.org/atl/usecases/)
- 4. Reimplement these case studies with alternative technologies (e.g., QVT)
- 5. Design a benchmark setup to systematically measure the energy consumption of model transformations
- 6. Analyze the quality of the conducted measurements using statistical tests
- 7. Draw conclusions from the measurment result
Betreuer: Sebastian Götz