The project will be built in several phases, starting with breakdowns per rack and server, starting with the direct energy consumption. The framework should then produce overviews of the daily, monthly, yearly consumption and climate footprint in a dynamic dashboard. Furthermore, we want to create updateable reports for compute jobs in the data center to make users aware about the power consumption and climate footprint. The model and framework will be generic to be reusable for other data centers.
Project partners
During this project, the participants will collaborate with colleagues from Hewlett Packard Enterprise and climate researcher Stefan Krottenthaler from University of Passau.
Skills
The participants need to have experience in software engineering and at least one programming language (C++, Java, Python). Basic experience in machine learning is beneficial. Participants should be comfortable with documenting their work and visualizing their results. The initial phase of the project will include literature research to find models and estimations for emission calculation. During the project, you will further your knowledge about software engineering in teams and programming, but also get deep insights into data center architecture and setup as well as climate impact prediction and analysis.
Contact
Please contact Tilmann Rabl for any questions.
Recommended reading
1. https://ccaf.io/cbeci/index/comparisons
2. https://www.goclimate.com/blog/the-carbon-footprint-of-servers/
3. In Computer Architecture, We Don't Change the Questions, We Change the Answers, Mark D. Hill, Keynote at Data Management on New Hardware (DaMoN) Workshop @SIGMOD, June 2022. Slides: pdf
4. https://ec.europa.eu/clima/eu-action/climate-strategies-targets/2050-long-term-strategy_en
5. https://de.wikipedia.org/wiki/GHG_Protocol
