L.M. ERICSSON LIMITED

The project aerOS aims at transparently utilising the resources on the edge-to-cloud computing continuum for enabling applications in an effective manner, incorporating multiple services deployed on such a path. Therefore, aerOS will establish the missing piece: a common meta operating system that follows a collaborative IoT-edge–cloud architecture supporting flexible deployments (e.g., federated or hierarchical), bringing tremendous benefits as it enables the distribution of intelligence and computation – including Artificial Intelligence (AI), Machine Learning (ML), and big data analytics – to achieve an optimal solution while satisfying the given constraints. The overarching goal of aerOS is to design and build a virtualized, platform-agnostic meta operating system for the IoT edge-cloud continuum. As a solution, to be executed on any Infrastructure Element within the IoT edge-cloud continuum – hence, independent from underlying hardware and operating system(s) – aerOS will: (i) deliver common virtualized services to enable orchestration, virtual communication (network-related programmable functions), and efficient support for frugal, explainable AI and creation of distributed data-driven applications; (ii) expose an API to be available anywhere and anytime (location-time independent), flexible, resilient and platform-agnostic; and (iii) include a set of infrastructural services and features addressing cybersecurity, trustworthiness and manageability. aerOS will: (a) use context-awareness to distribute software task (application) execution requests; (b) support intelligence as close to the events as possible; (c) support execution of services using “abstract resources” (e.g., virtual machines, containers) connected through a smart network infrastructure; (d) allocate and orchestrate abstract resources, responsible for executing service chain(s) and (e) support for scalable data autonomy.

EC contribution:

EUR 416k

Start:

2022-09-01

End:

2025-10-31

Status:

SIGNED

Scheme:

HORIZON-RIA

Call:

HORIZON-CL4-2021-DATA-01

View on CORDIS (DOI 10.3030/101069732)

GENerative and connected intelligence for 6G Open ManagemEnt (GENOME) research and training programme aims at investigating, developing/fine-tuning and validating large language models (LLMs) for the autonomous management and orchestration of various network technological domains (O-RAN, edge, cloud) through intent-based networking and task-agnostic network functions. It also targets the establishment of a connected intelligence framework for the coordination between these functions via on-the-fly protocol learning to mitigate conflicting management decisions and maximize utilization. Network configuration will also be generated automatically through innovative neuro-symbolic AI algorithms. Moreover, the transparency and resilience of the AI models is researched. All the developed components will be validated via both simulation and testbed experiments, and integrated into a final proof-of-concept. Besides, a detailed training plan is designed including schools, industrial days and soft-skills courses. It involves industrial partners to ensure that career perspectives of ESRs will be significantly increased by participation in the project. High quality dissemination and exploitation activities are also considered.

EC contribution:

EUR 341k

Start:

2026-01-01

End:

2029-12-31

Status:

SIGNED

Scheme:

HORIZON-TMA-MSCA-DN

Call:

HORIZON-MSCA-2024-DN-01