Press release: New solutions for cloud computing management
No. 144 - 18.10.2022
University of Göttingen and GWDG coordinate Europe-wide joint project with 13 partners
The cloud computing industry has grown massively over the past ten years. This has opened up many new areas at the same time as the demands on the hardware and performance of cloud applications have increased. Modern cloud applications have also become increasingly complex, as they are usually based on distributed computer systems, with individual components that must run with high availability. How such disparate systems can be combined and unified into centrally controlled computing clusters, and how scheduling decisions might proceed, is the focus of a new Europe-wide collaborative project coordinated by the University of Göttingen with the participation of the Gesellschaft für wissenschaftliche Datenverarbeitung mbH Göttingen (GWDG).
The Device-Edge Cloud Intelligent Collaboration Framework (DECICE) project aims to develop a cloud management framework for optimizing applications: in a heterogeneous system, jobs are to be automatically and adaptively allocated to the most suitable resources. A digital twin of the system that mirrors the entire original system is used for holistic monitoring. An artificial intelligence (AI)-based scheduler then makes decisions about the placement of jobs and data, as well as rescheduling jobs if they need to be adapted to system changes. In addition, a virtual training environment will be provided that generates test data for training machine learning models and investigating “what-if” scenarios.
The DECICE project involves a total of 13 partners in Germany, Italy, Sweden, Austria, Turkey and the United Kingdom. The European Union has funded the project through the Horizon Europe program for three years with a total of around 5.9 million euros. Of this, the University of Göttingen and the GWDG will receive around 1.37 million euros.
Professor Julian Kunkel
University of Göttingen
Faculty of Mathematics and Computer Science
Institute for Computer Science
Tel: +49 (0)551 39-30144