Sub-project Reference No.: NextGenerationEU [UNICO-5G I+D/FREE6G-Reg Edge (TSI-063000-2021-144)]
FREE6G-RegEdge aims at providing an integrated orchestration and management of communication, storage and computational resources. To this end, FREE6G-RegEdge aims to optimize the functionality of the MEC system and the Network Slicing Management via a hierarchy of analytic and decision engines, proposing novel ML-based approaches for distributed representation based on multi-objective optimization. Moreover, FREE6G-RegEdge envisages a new approach based on the Elastic Edge Computing notion of an evolved MEC system which allows application functions to by disaggregated and dynamically migrated across the edge and also from the RAN to the core cloud and vice-versa. This implies the need to support advanced intra-DC and inter-DC load balancing scenarios addressing optimized distribution of latency budgets. The ultimate goal is to achieve a zero perceived latency primarily in MEC applications requiring smart connectivity. Finally, dynamic slicing approaches will also be explored, to trigger predictive slice reconfiguration across domains.
FREE6G-RegEdge NW Architecture
-
The specific technical objectives of FREE6G-RegEdge are:
-
Objective 1: Deployment of an Elastic Edge Computing paradigm with Cloud-Native technologies.
i) Extend the MEC system with the capability to disaggregate and place MEC app functions to any Edge Data centre, and steer UE traffic towards any MEC host, for inter-DC load balancing.
ii) Study dynamic Virtual Network Embedding, determining the optimal placement of application functions at Edge Data-Centres, considering compute, networking and storage constraints.
-
Objective 2: Deliver a Self-driven infrastructure with pervasive, ML-driven control.
i) Explore Representation learning algorithms for the Analytic Engines to build a knowledge graph of network Slices and MEC applications, representing network context.
ii) Propose distributed techniques based on Embedding Propagation to derive the network context representations independently in each Edge node.
iii) Explore multi-objective optimization and multi-agent Deep Reinforcement Learning for the decision engines, planning and reacting to current context based on policy requirements.
-
Objective 3: implementation and proof-of-concept of the FREE6G-RegEdge solutions. One Proof-of-Concept (PoC) case will be implemented under the FREE6G-RegEdge framework to demonstrate the deployment of FREE6G-RegEdge elastic MEC platform and zero-touch network slice reconfiguration to meet stringent 6G requirements.
-
Objective 4: Dissemination, standardization, and exploitation of FREE6G-RegEdge.
i) Dissemination to relevant industrial and academic communities.
ii) Dissemination of results at EC related initiatives.
iii) Communication outreach to all stakeholders including the general public.
iv) Contributions to standardization bodies.
v) Cross fertilization with and contributions to relevant European WGs.
vi) Contributions to top-tier conferences and journals.
vii) Generation of intellectual property.
