An AI Framework for Fostering 6G towards Energy Efficiency

The sixth generation of radio mobile networks (6G) is expected to employ Artificial Intelligence (AI) to support this new network evolution. Along with the enabling technologies of its predecessor, namely edge computing and network slicing, AI will realize intelligent communication and networking, but at a high risk of severely increasing the demands on computing resources and the related carbon footprint. In order to avoid such a curse, this paper proposes an AI framework allowing to achieve tight proportionality between the workload produced by vertical applications and network slices, and the energy consumption induced in the infrastructure, and enabling beyond 5G stakeholders to become part of a green business model. The paper outlines a high-level architecture able to support the framework, as well as the role of the algorithms composing it. Results show that the proposed framework allows for energy savings over 100% while fulfilling the application requirements.