Advanced Edge Computing Architecture for AI-Driven Automation and Slicing in Beyond 5G

This paper proposes an open, modular, and distributed architecture for the edge-to-cloud compute continuum, in synergy with the beyond 5G and 6G architectural trends. The proposed design facilitates the lifecycle management of cloud-native services, as well as Artificial Intelligence (AI) pipelines developed for the optimization of complex and heterogeneous edge-enabled environments, while supporting a suite of novel mechanisms to ensure the trustworthiness of AI solutions. Finally, the proposed architecture introduces a potential extension of the network slicing paradigm for the joint provisioning of communication and computing resources for applications running at the edge.