Objective 1: Design an open and secure edge computing architecture to efficiently support highly demanding XR and IoT-driven applications in terms of computation and connectivity, relying on an edge-cloud continuum that fosters virtualisation of massive heterogeneous edge computing resources across a multi-access RAN.

This objective covers the specification and design of the VERGE system architecture for next-generation edge-cloud computing able to deliver the high data rate and latency demands of B5G AI-enabled XR applications and use cases. It will be sustained on a multi-domain edge-cloud computation continuum distributed across a multi-access RAN encompassing both fixed RAN nodes and moving relays (e.g., at drones, trams, etc.).

Objective 2: Design and build the "edge for AI" concept that provides the infrastructure, mechanisms and programming tools to enable the development, deployment and orchestration of massively distributed AI processes across heterogeneous computing and memory resources.

This objective targets the development of the components of the "edge for AI" framework of the VERGE system architecture. This framework should provide the overlay/abstraction capabilities to flexibly decompose and recompose heterogeneous computing and memory resources at a multi-edge architecture and to flexibly distribute the different computation tasks, under strict latency, energy efficiency and performance budgets. For this purpose, a novel software programming abstraction layer will be developed to fully exploit the capabilities of HW-accelerated platforms for ultra-high performance computation.

Objective 3: Design an "AI for edge" framework that encompasses cutting-edge AI solutions for managing and orchestrating the edge computing and the RAN resources towards an optimum  performance that satisfies the highly demanding applications requiring extremely  low latencies and/or very high-capacity justifying edge processing and computing.

The "AI for edge" framework will include the algorithmic solutions for achieving an optimized operation of the VERGE system. These solutions will rely on the monitoring capabilities available at the "edge for AI", which will provide different metrics characterising the behaviour of the compute continuum resources and the network infrastructure composed of the RAN and the core network domains. Based on this, this objective intends to develop advanced AI solutions that manage and orchestrate the underlying physical, network and compute resources.

Objective 4: Develop tools that ensure the security, privacy and trustworthiness of the VERGE system.

The VERGE system will provide a highly distributed framework for executing applications across different domains of the compute continuum and for decentralizing AI-based solutions. This leads to different concerns in relation to security, privacy and trust that VERGE intends to address. First, AI systems should be resistant to adversarial attacks that can affect the inference and training phases. Second, AI systems are subject to privacy related threats affecting data containing private and sensitive information. Third, the success of AI depends on building solutions that humans can trust, which involves aspects such as robustness, explainability or safety of AI-based solutions.

Objective 5: Showcase the VERGE solutions by means of Proof of Concept (PoC) demonstrations.

The VERGE system will provide a highly distributed framework for executing applications across different domains of the compute continuum and for decentralizing AI-based solutions. This leads to different concerns in relation to security, privacy and trust that VERGE intends to address. First, AI systems should be resistant to adversarial attacks that can affect the inference and training phases. Second, AI systems are subject to privacy related threats affecting data containing private and sensitive information. Third, the success of AI depends on building solutions that humans can trust, which involves aspects such as robustness, explainability or safety of AI-based solutions.

Objective 6 Carry out extensive dissemination, standardisation and exploitation activities.

VERGE project will disseminate and communicate its achievements through multiple channels.

The aim is to reach both specialised and non-specialised audiences and to create awareness of VERGE concepts and benefits within the society and the relevant stakeholders. Participation in working groups of the SNS-JU will also be used as a means to communicate the project achievements and to establish liaisons with other projects for identifying synergies and consolidating concepts towards future B5G/6G networks. The project also targets the  contribution to relevant standardization bodies (e.g., ETSI MEC, ETSI ZSM, 3GPP). Moreover, the project intends to create a VERGE open dataspace for enabling the public access to the different datasets generated by the project. VERGE also intends to achieve an industrial and academic exploitation of the different concepts and results obtained by the project.