Periodic Reporting for period 2 - GREENEDGE (Taming the environmental impact of mobile networks through GREEN EDGE computing platforms)
Reporting period: 2023-03-01 to 2025-02-28
Mobile applications and services along with fifth generation (5G) and beyond 5G networks have led to a digital revolution, enabling ubiquitous and enhanced broadband services, smart/autonomous vehicles, intelligent transport, and complex human machine interactions. Hence, our connected society generates a vast amount of data, which has to be transmitted, processed and analysed, often at runtime. Such a data surge demands a massive use of computing services based on artificial intelligence (AI) and machine learning (ML) algorithms at the network edge, which have a large, unsustainable, carbon footprint.
GREENEDGE aimed at mitigating the carbon footprint of Information and Communication Technology (ICT), including mobile networks and AI-based computing systems, being aligned with mitigation policies for climate change at EU and UN levels (reduction of carbon emissions and achievement of carbon neutrality of human activities by 2050). Sustainable development of ICT is essential for achieving these objectives. In fact, ICT was used extensively during the last decade and it is estimated to be consuming around 20% of electricity worldwide by 2030.
GREENEDGE tamed the growing carbon footprint of ML/AI based edge computing services with a twofold approach:
- First, renewable energy resources were utilised to sustain the communication and computing tasks that run at the mobile network edge.
- Second, green-by-design computing and communication paradigms were developed, to wisely exploit the available communication, AI/ML/computing and energy resources across the mobile edge network. Particular attention was paid to distributed computing approaches (federated learning).
In conclusion, the project has successfully achieved the planned objectives, developing new technologies for green edge computing and training highly qualified researchers through six training schools, local- and secondment-based actions. One final workshop was organised and 40 scientific publications were published. Among the major achievements, we underline the development of prototype edge nodes powered by light and with very high energy efficiency. Other notable results include new algorithms and theories, with strategies that led to substantial energy savings against the state-of-the-art. The project set a solid basis for future developments on green energy ICT systems, and contributed to increasing the awareness around green ICT.
GREENEDGE aimed at mitigating the carbon footprint of Information and Communication Technology (ICT), including mobile networks and AI-based computing systems, being aligned with mitigation policies for climate change at EU and UN levels (reduction of carbon emissions and achievement of carbon neutrality of human activities by 2050). Sustainable development of ICT is essential for achieving these objectives. In fact, ICT was used extensively during the last decade and it is estimated to be consuming around 20% of electricity worldwide by 2030.
GREENEDGE tamed the growing carbon footprint of ML/AI based edge computing services with a twofold approach:
- First, renewable energy resources were utilised to sustain the communication and computing tasks that run at the mobile network edge.
- Second, green-by-design computing and communication paradigms were developed, to wisely exploit the available communication, AI/ML/computing and energy resources across the mobile edge network. Particular attention was paid to distributed computing approaches (federated learning).
In conclusion, the project has successfully achieved the planned objectives, developing new technologies for green edge computing and training highly qualified researchers through six training schools, local- and secondment-based actions. One final workshop was organised and 40 scientific publications were published. Among the major achievements, we underline the development of prototype edge nodes powered by light and with very high energy efficiency. Other notable results include new algorithms and theories, with strategies that led to substantial energy savings against the state-of-the-art. The project set a solid basis for future developments on green energy ICT systems, and contributed to increasing the awareness around green ICT.
Since the project started, the GREENEDGE consortium has actively run the training network, by pursuing all the objectives stated in the Annex 1.
After four years, the GREENEDGE LinkedIn account counted 303 connections and 479 followers and our posts received more than 46798 views (updated April 10, 2025).
Following the GA Annex 1, six training schools, two special sessions and one final workshop at international conferences were organised, covering all the planned technical and soft-skill trainings. The 2nd training school was co-organized as a PhD summer school open to a wider international community of PhD researchers, co-organizing it with EU MINTS ITN. Including GRENEDGE and MINTS fellows, this school has attracted more than 100 PhD researchers from across Europe.
The GREENEDGE Youtube channel was populated with 40 videos, with generalist content and with more specific technical sessions by our fellows. Five videos, presented in a dedicated playlist introduce the project, our goals, our research fellows and seniors, creating awareness around GREENEDGE objectives and actions. Short presentation videos were published for all the ESRs to introduce who they are, their research topic and aspirations. Following these, additional videos created by the ESRs were produced and uploaded, by delving more into the technical details of their specific individual projects. Finally, 12 selected lessons were uploaded. At the project end, the YouTube channel had 66 subscribers and received 6292 views.
All the schools and training events were implemented, as per the description of work in the Annex 1.
Forty scientific publications were published. Among the results, several major improvements were obtained, by developing various tools for measuring the energy consumption of edge devices and more energy efficient algorithms for training edge computing software at the network edge. In some cases, substantial energy reduction was attained (ranging from 40 to 82% depending on the specific problem and scenario).
Among the main impacts, we cite:
- New ultra-low power solutions for edge sensing networks (exploitation into commercial products): a new technology for IoT/edge networks was created to manage edge sensing networks with very low duty cycle and very high energy efficiency, capable of lasting several years without human intervention/battery replacements, etc. This technology was designed, implemented into Worldsensing products, validated by ESR13, and exploited within their products, becoming one of the key technological assets of the company.
- New batteryless light-based devices (prototypes): KUL and OULU collaborated (ESR09 and ESR10) in the creation of a new breed of batteryless edge sensors. These edge devices consume extremely low energy and are solely operated by solar panels. They turn on as soon as they receive environmental energy and communicate via Light-based technology (visible light communication). So light is used for powering up the nodes and for sending and receiving data to/from them.
- Participation in new competitive calls: The recent EU SuperIoT project sparkled from the collaboration and technical work that were accomplished in GREENEDGE.
- Patents: our beneficiary CEA obtained a patent thanks to the activity carried out by ESR11 about energy efficient methods for detecting intrusions in computer networks based on machine learning.
After four years, the GREENEDGE LinkedIn account counted 303 connections and 479 followers and our posts received more than 46798 views (updated April 10, 2025).
Following the GA Annex 1, six training schools, two special sessions and one final workshop at international conferences were organised, covering all the planned technical and soft-skill trainings. The 2nd training school was co-organized as a PhD summer school open to a wider international community of PhD researchers, co-organizing it with EU MINTS ITN. Including GRENEDGE and MINTS fellows, this school has attracted more than 100 PhD researchers from across Europe.
The GREENEDGE Youtube channel was populated with 40 videos, with generalist content and with more specific technical sessions by our fellows. Five videos, presented in a dedicated playlist introduce the project, our goals, our research fellows and seniors, creating awareness around GREENEDGE objectives and actions. Short presentation videos were published for all the ESRs to introduce who they are, their research topic and aspirations. Following these, additional videos created by the ESRs were produced and uploaded, by delving more into the technical details of their specific individual projects. Finally, 12 selected lessons were uploaded. At the project end, the YouTube channel had 66 subscribers and received 6292 views.
All the schools and training events were implemented, as per the description of work in the Annex 1.
Forty scientific publications were published. Among the results, several major improvements were obtained, by developing various tools for measuring the energy consumption of edge devices and more energy efficient algorithms for training edge computing software at the network edge. In some cases, substantial energy reduction was attained (ranging from 40 to 82% depending on the specific problem and scenario).
Among the main impacts, we cite:
- New ultra-low power solutions for edge sensing networks (exploitation into commercial products): a new technology for IoT/edge networks was created to manage edge sensing networks with very low duty cycle and very high energy efficiency, capable of lasting several years without human intervention/battery replacements, etc. This technology was designed, implemented into Worldsensing products, validated by ESR13, and exploited within their products, becoming one of the key technological assets of the company.
- New batteryless light-based devices (prototypes): KUL and OULU collaborated (ESR09 and ESR10) in the creation of a new breed of batteryless edge sensors. These edge devices consume extremely low energy and are solely operated by solar panels. They turn on as soon as they receive environmental energy and communicate via Light-based technology (visible light communication). So light is used for powering up the nodes and for sending and receiving data to/from them.
- Participation in new competitive calls: The recent EU SuperIoT project sparkled from the collaboration and technical work that were accomplished in GREENEDGE.
- Patents: our beneficiary CEA obtained a patent thanks to the activity carried out by ESR11 about energy efficient methods for detecting intrusions in computer networks based on machine learning.
The GREENEDGE project had three original aspects that differentiated it from the state-of-the-art and other doctoral programs.
1. it was set to devise highly energy efficient communication and computing techniques with a holistic look at the underlying computing/learning applications, alongside an intelligent network management spanning across all layers.
2. it accounted for any distributed energy generation and harvesting technologies, and energy storage hardware that the network may exploit.
3. it contributed to the EU Green Deal by introducing metrics and methodologies to measure the energy efficiency and the environmental footprint of ICT (with focus on computing and communication) by devising energy management techniques to reduce its carbon footprint.
GREENEDGE fellows were trained in a cross-disciplinary field of knowledge including: energy generation and storage, communication/coding theory, mobile network design and management, machine learning and mathematical optimisation, and distributed computing. The fellows worked in a multicultural environment, building a solid basis to enable their future transnational research, and exploiting synergies among the international partners of the consortium.
Among the main achievement, we cite:
- New ultra-low power solutions for edge sensing networks were developed and integrated into products.
- New batteryless light-based devices solely operated by solar panels were demonstrated/prototyped.
GREENEDGE technology had a strong and positive impact on GREENEDGE beneficiaries, consolidating their leading position in green ICT systems.
1. it was set to devise highly energy efficient communication and computing techniques with a holistic look at the underlying computing/learning applications, alongside an intelligent network management spanning across all layers.
2. it accounted for any distributed energy generation and harvesting technologies, and energy storage hardware that the network may exploit.
3. it contributed to the EU Green Deal by introducing metrics and methodologies to measure the energy efficiency and the environmental footprint of ICT (with focus on computing and communication) by devising energy management techniques to reduce its carbon footprint.
GREENEDGE fellows were trained in a cross-disciplinary field of knowledge including: energy generation and storage, communication/coding theory, mobile network design and management, machine learning and mathematical optimisation, and distributed computing. The fellows worked in a multicultural environment, building a solid basis to enable their future transnational research, and exploiting synergies among the international partners of the consortium.
Among the main achievement, we cite:
- New ultra-low power solutions for edge sensing networks were developed and integrated into products.
- New batteryless light-based devices solely operated by solar panels were demonstrated/prototyped.
GREENEDGE technology had a strong and positive impact on GREENEDGE beneficiaries, consolidating their leading position in green ICT systems.