Periodic Reporting for period 1 - ECO-eNET (Efficient Confluent Edge Networks)
Periodo di rendicontazione: 2024-01-01 al 2025-06-30
The ECO-eNET project addresses one of the central challenges of 6G: how to deliver ultra-fast, reliable and energy-efficient connectivity to everyone, everywhere, while keeping costs and environmental impact under control. Future 6G networks will need to support highly demanding services such as immersive communication, telepresence and critical machine-to-machine applications. However, today’s infrastructure often struggles with deployment costs, energy consumption and service gaps in rural or hard-to-reach areas.
To respond to this challenge, ECO-eNET introduces the concept of confluent networking. This novel approach integrates radio fixed wireless, free-space optics and fibre transmission into flexible mesh topologies. This allows fibre-like performance even in locations where fibre deployment is difficult or uneconomical, while ensuring low latency and very low energy per bit. The project pursues four main scientific objectives: (i) the development of advanced transmission technologies based on plasmonics, photonics and fibre sensing; (ii) the creation of intelligent orchestration methods that unify the control of mobile, optical and compute resources; (iii) the integration and demonstration of these technologies in realistic proof-of-concept environments; and (iv) the dissemination and standardisation of results to maximise long-term impact.
Through this approach, ECO-eNET seeks to establish confluent networks as a scalable and energy-aware platform for 6G, directly supporting EU priorities for climate action, digital inclusion, and industrial leadership.
To respond to this challenge, ECO-eNET introduces the concept of confluent networking. This novel approach integrates radio fixed wireless, free-space optics and fibre transmission into flexible mesh topologies. This allows fibre-like performance even in locations where fibre deployment is difficult or uneconomical, while ensuring low latency and very low energy per bit. The project pursues four main scientific objectives: (i) the development of advanced transmission technologies based on plasmonics, photonics and fibre sensing; (ii) the creation of intelligent orchestration methods that unify the control of mobile, optical and compute resources; (iii) the integration and demonstration of these technologies in realistic proof-of-concept environments; and (iv) the dissemination and standardisation of results to maximise long-term impact.
Through this approach, ECO-eNET seeks to establish confluent networks as a scalable and energy-aware platform for 6G, directly supporting EU priorities for climate action, digital inclusion, and industrial leadership.
During the first 18 months, ECO-eNET has made substantial progress across all its objectives. In transmission technology, plasmonic sub-THz links have been designed to target one terabit per second over one kilometre at an energy consumption below 40 picojoules per bit. Simulations confirmed feasibility, showing a signal-to-noise ratio of around 20 decibels. Optical wireless links have been advanced through the development of adaptive optics based on photonic lanterns and photonic integrated circuits, with the first prototypes already fabricated. Fibre sensing has been implemented in the form of a Telemetry-as-a-Service platform, which achieved an F1 score of approximately 93 percent in anomaly detection. The project also carried out confluent xhaul experiments that demonstrated the coexistence of multi-gigabit radio-over-fibre signals and 400 gigabit coherent optical channels in metro and passive optical networks, validated in a live trial on the HEAnet production network.
On the orchestration side, a unified monitoring framework has been created, spanning radio access, core and transport domains and integrating energy as well as performance metrics through open-source tools such as Prometheus and Grafana. AI-based orchestration methods have been developed for integrated access and backhaul networks, centralised RANs and optical fronthaul switching, with initial results already disseminated in scientific publications. Some preliminary work on federated AI orchestration has explored how to balance computing between the edge and the central cloud and has applied neural architectures inspired by natural language processing to quality-of-transmission estimation in optical networks.
The project has also achieved strong visibility, with 27 peer-reviewed journal articles and 20 international conference presentations at major venues such as OFC, ECOC, EuCNC-6G Summit and Fyuz. ECO-eNET advanced from Technology Readiness Level 2 at project start to TRL 3 by mid-term, on track to reach TRL 4 by project completion.
On the orchestration side, a unified monitoring framework has been created, spanning radio access, core and transport domains and integrating energy as well as performance metrics through open-source tools such as Prometheus and Grafana. AI-based orchestration methods have been developed for integrated access and backhaul networks, centralised RANs and optical fronthaul switching, with initial results already disseminated in scientific publications. Some preliminary work on federated AI orchestration has explored how to balance computing between the edge and the central cloud and has applied neural architectures inspired by natural language processing to quality-of-transmission estimation in optical networks.
The project has also achieved strong visibility, with 27 peer-reviewed journal articles and 20 international conference presentations at major venues such as OFC, ECOC, EuCNC-6G Summit and Fyuz. ECO-eNET advanced from Technology Readiness Level 2 at project start to TRL 3 by mid-term, on track to reach TRL 4 by project completion.
ECO-eNET will be demonstrating the integration of sub-THz, free-space optical and fibre technologies within a unified confluent networking framework combined with AI-enhanced orchestration. Whereas these media have traditionally been studied in isolation, the project aims to show how they can be used together to deliver excellent performance in terms of capacity, energy efficiency and latency.
The innovations developed go well beyond current practice. Plasmonic transceivers overcome the bandwidth limitations of conventional electronic radios and promise to cut energy use per bit by orders of magnitude. Photonic adaptive optics enable transparent fibre-to-fibre free-space links without the need for heavy and power-hungry digital signal processing. Fibre-sensing-driven orchestration turns passive infrastructure into a distributed sensor network, reducing the need for on-site maintenance and enabling predictive interventions. Finally, end-to-end orchestration across wireless, optical and compute domains makes possible dynamic service provisioning and intelligent slice management.
These advances are already showing industrial relevance. Project partners are evaluating the integration of fibre sensing into commercial network monitoring platforms, the embedding of confluent orchestration into Open RAN systems, and the upgrade of free-space optical terminals with adaptive optics. Time-to-market for several of these innovations is estimated at between two and five years. To achieve complete uptake, further work will be needed on large-scale demonstrations in live networks, testing with end users, developing business models, and engagement in standardisation and regulatory processes.
The innovations developed go well beyond current practice. Plasmonic transceivers overcome the bandwidth limitations of conventional electronic radios and promise to cut energy use per bit by orders of magnitude. Photonic adaptive optics enable transparent fibre-to-fibre free-space links without the need for heavy and power-hungry digital signal processing. Fibre-sensing-driven orchestration turns passive infrastructure into a distributed sensor network, reducing the need for on-site maintenance and enabling predictive interventions. Finally, end-to-end orchestration across wireless, optical and compute domains makes possible dynamic service provisioning and intelligent slice management.
These advances are already showing industrial relevance. Project partners are evaluating the integration of fibre sensing into commercial network monitoring platforms, the embedding of confluent orchestration into Open RAN systems, and the upgrade of free-space optical terminals with adaptive optics. Time-to-market for several of these innovations is estimated at between two and five years. To achieve complete uptake, further work will be needed on large-scale demonstrations in live networks, testing with end users, developing business models, and engagement in standardisation and regulatory processes.