Energy-efficient Computing and Green Edge Grid: Data-driven Modelling (engl.)

To limit carbon emissions, recent research shows that utilizing photovoltaic energy sources rather than fossil fuels to produce electricity reduces the final energy demand by up to 40%. Therefore, this course aims to learn about the methods to design and devise strategies for reaching a more zero carbon emission environment.
Firstly, we learn about household solar power production predictions by applying machine learning models to a specific dataset. Households are at the Edge of the smart grid network; therefore, we utilize the available data traces and several categories of machine learning models, such as boosting and neural networks, to walk towards a greener environment.
On the other hand, one solution to help infrastructure providers, such as Cloud or Edge, achieve a greener environment is to utilize optimized orchestration strategies for distributing the user’s services. Hence, this course also covers working with Docker containers, microservice orchestrators such as Kubernetes, and developing a scheduling strategy with the help of a Kubernetes Python Client. This section of the course targets strategies to lower resource utilization and energy consumption.
We use the libraries and APIs available in Python for all the aforementioned steps.

***
Requirements: –
Credit Points (ECTS): 1, grading possible