5G, AI, passive cooling and integration combine to reduce network energy consumption
New technologies are dramatically improving the energy efficiency of mobile networks, while reducing their greenhouse gas emissions. That was one of the key takeaways from a recent GSMA webinar exploring the impact of 5G-Advanced on energy efficiency.
In Europe, for example, mobile network operators (MNOs) have halved their operational emissions in just three years, according to George Kamiya, senior manager of climate action at the GSMA, who outlined the key findings from the GSMA’s new Mobile Net Zero 2024 report. “Globally, MNOs’ operational emissions were down six percent between 2019 and 2022, and this is despite surging demand for data connectivity,” he explained. “Europe is leading away with the remarkable 50 per cent reduction in just three years. North America, Latin America and MENA (Middle East and North Africa) also saw strong reductions of over 20 percent.”
George Kamiya explained that these reductions were achieved through a combination of energy efficiency and greater usage of renewable energy (which can also now cost less than conventional sources). “Energy efficiency, electrification of fleets and generators and renewable energy accounted for more than 80 percent of the reductions,” he added.
As older generations of mobile networks are retired, there should be further energy savings. “An operator in the UK reported that 3G represents a third of its mobile network energy use, but only 0.6 per cent of data traffic and seven percent of voice traffic,” George Kamiya said. “So, lots of potential there in terms of reducing energy use in the near term.”
Cutting the complexity with AI
At the same time, artificial intelligence (AI) is increasingly being employed in this data-intensive and complex aspect of network management. Operators may have to make thousands, maybe even millions of decisions, each day to optimise their energy usage. While that would be impossible with a manual approach, AI can make it feasible.
Telenor is working with Ericsson to develop a “multi-purpose algorithm that we can use to optimise all the parts of the network,” explained Johannes Bjelland, program director networks, at Telenor Research & Innovation, noting that AI can help operators maximise the impact of sleep mode and other energy saving mechanisms. “You don’t go in and tell the algorithm what to do. The idea is to use the simulator approach to let the algorithm learn the best policy to actually save power,” he added. “We’re testing them on one frequency band now and looking at around four percent power efficiency, on top of the regular sensor-led approach that we use today, so it’s promising.”
AI can achieve energy savings across the time domain, the frequency domain and the spatial domain, added Emanuel Kolta, lead analyst, network sustainability & innovation at GSMA intelligence. “The time domain is probably the easiest,” and an area of focus for most solutions right now, he noted. In the frequency domain, AI can be used to select the right spectrum band for the right use case, while in the spatial domain, energy usage can be dialled up and down geographically depending on demand.
Emanuel Kolta said the spatial domain is the most complex because of the way in which end-user devices move from place to place. “I actually saw a private network recently, and it was very impressive because the antenna system could actually change the coverage area,” he recalled. “During the day, it was a focusing onto the office area and ….[then] put more focus into the canteen during the lunch break.” By anticipating how demand will change, AI can ensure network resources are deployed most efficiently.
Simpler, cheaper and less power-hungry
Emanuel Kolta also pointed to a positive impact from the simplification of cell sites, greater integration of hardware and advances in cooling technologies. He highlighted how cell sites, for example, are becoming far more streamlined, as they increasingly share baseband units and make greater use of fibre connections. “We see this very rapid change, which is heavily impacting energy efficiency” because don’t need as many different power modules, he added.
In particular, Emanuel Kolta underlined the importance of advanced cooling solutions, as many operators spend more than 50 percent of their total energy consumption at a site on passive infrastructure and cooling. He explained that new passive solutions, using new materials, new design and new components, are “getting rid of excess heat way quicker than active solutions and also way cheaper.”
Although operators have introduced many energy-savings features in their networks, there is still scope for improvement. Orange is developing a solution that could reduce a 5G network’s energy consumption to almost zero when it isn’t transmitting traffic, explained Jean-Marie Chaufray, head of RAN Benchmark and Green at Orange. “Now there are very promising technology to do that,” he explained. “Through energy-saving features and also through what we call the zero bit zero watts (kind of an extreme deep sleep mode for the 5G antenna), so that when there is no 5G traffic, we can reduce, not to zero, but close to zero, the consumption of the 5G antennas.”
Join the GSMA 5G Futures Community
Join the GSMA 5G Futures Community to stay up to date with the latest industry updates, upcoming GSMA events and be the first to access the latest releases from the GSMA. Register to join the community and be part of the conversation today.