12/12/2024 | News release | Archived content
Almost every aspect of modern life is underpinned by mobile data communications, from banking to healthcare to our social networks. Our society generates a vast amount of data, which must be transmitted, processed and analysed, often in an online fashion. To reduce congestion on the network, computing services are increasingly being pushed to users' devices and mobile base stations, a strategy known as edge computing. This is further increasing the energy needed to support our online lives.
To address this, the GREENEDGE project is supporting the development of energy-efficient computing systems for mobile edge technology. Funded through the EU's Marie Skłodowska-Curie Actions programme, GREENEDGE enlisted 15 early-stage researchers from across the world to follow a two-pronged approach: create greener communication and greener-by-design computing.
Energy-hungry networks
Studies show the power consumption of data networks is rising dramatically, and now rivals that of the aviation sector. These energy-hungry networks rely heavily on computing services powered by both AI and ML algorithms.
With the advent of 5G and new AI-based services, these algorithms will shift to the edge of the mobile wireless network, and their global carbon footprint will significantly increase. The GREENEDGE project recognised the urgency of this issue and proposed innovative research to address it.
"Technology is becoming much more complex," explains GREENEDGE project coordinator Michele Rossi, a professor of Information Engineering at the University of Padova in Italy. "And this complexity requires significantly more communication and more computation. GREENEDGE aims to replace today's technology with something that is more efficient from an energetic perspective."
Information innovation
The project's researchers worked on developing new, energy-efficient algorithms for training and running AI models within networks, as well as designing and implementing hardware solutions that prioritised low energy consumption. They also explored the feasibility of using renewable energy sources, such as solar panels, to power network infrastructure. Finally, they developed cutting-edge AI algorithms for edge computing that were significantly more energy-efficient compared to existing models.
Students within the network trained in technical aspects of the project at several research institutions across Europe and collaborated with industry partners such as telecom operators, internet of things specialists, and smart city developers. At the same time, they were also equipped with soft skills during workshops and training sessions, such as marketing and project management, to prepare them to become future leaders in the field of sustainable information technology.
As Rossi highlights: "Our researchers are under more pressure than standard PhD students. With more training, more responsibility, more deliverables and more commitments dictated by the project timeline. By the end, we see a professional - and I believe this is a big advantage for them."
So far, the GREENEDGE project has produced some groundbreaking results. They've achieved a 43 % reduction in energy consumption on edge devices to enhance speed and battery life, and developed smarter AI training that reduces data transfer and power usage. Additionally, they've proposed an energy-saving traffic forecasting system for smart cities and algorithms to optimise battery life for AI tasks.
Most notably, GREENEDGE partners have received further EU funding for SUPERIOT, a project which aims to develop a prototype for a sustainable, battery-free internet of things device. These small units can be installed discreetly on walls and powered by printable micro solar panels using photovoltaic or indoor light, making them highly energy-efficient. In the future, these nodes could even be printed on biodegradable substrates that dissolve after use, making them even more eco-friendly.
The future of ICT
These impressive results have been showcased in several peer-reviewed journals and at six conventions around the world, including the 2023 IEEE Global Communications Conference in Kuala Lumpur. In addition, the project co-organised two workshops on related topics at the 2024 IEEE PerCom conference to expand the community of researchers working on sustainable information technology.
But perhaps GREENEDGE's most significant accomplishment lies beyond its immediate achievements. By nurturing the next generation of green computing experts, they're equipping the innovators of the future to tackle the data deluge of our ever-expanding online presence and the surge of connected devices to come.