As a general estimation, telecom networks currently consume 1.5 times more power than global data centres; over the next decade network traffic is expected to grow tenfold with the explosion of technologies such as 5G, the Internet of things, high-bandwidth data centre networks and private mobile networks. According to the GSMA, the contribution of the telecom sector to global carbon emissions will rise from about 0.4% to 14% by 2040 if left unchecked. The importance of sustainability in IT, particularly in mobile networks, cannot be overstated.
I recently had the opportunity to discuss green networks and carbon reduction with Jeni Panhorst, vice president of Intel’s network and edge group. She explained that embodied carbon in products and operational carbon in processes both need to be addressed; there are numerous ways to do this, from using less power-hungry equipment to creating more energy efficient operations.
Solutions for lowering carbon emissions tend to fall into two categories: some focus on greater efficiency, others on using more renewable energy sources. For Intel, marketing itself as a solutions provider means having credibility on both fronts, not only in lowering its own operational and embodied carbon but also in delivering products that help customers cut their carbon footprint.
In April 2022, Intel announced an ambitious plan to hit two milestones: a 30% reduction in greenhouse gas emissions by 2030 and net zero emissions by 2040. A major part of its strategy to hit these targets is to increase reliance on renewables and water sustainability; this focus led to Intel being placed by financial news outlet Barron’s as number-one on its list of America’s Most Sustainable Companies.
For telecom operator customers, decreasing energy consumption in the network means understanding where and how it’s being used. Currently, 60% of total network energy is consumed in mobile networks and 40% in fixed networks. This is a growing problem for mobile networks — every time a new generation of technology is introduced, there’s a corresponding rise in energy consumption, based on life cycle data analysis. Deployment of 5G increases energy use because the radio antennas need more power and capacity.
This means a disproportionate amount of energy use in networks occurs in the final stage of service delivery, so reductions in the radio access network could potentially have the biggest impact. This energy use is because the radio access network in 4G or 5G is typically configured to be “always on” — fully powered for uptime as well as resilience obligations, and designed to support the highest possible data throughput. So the question is: can energy reductions be delivered with more intelligence in traffic monitoring, ensuring that there are no dropped packets or frames, no downtime and no compromising of signalling or radio availability?
Although traffic on mobile networks can fluctuate dynamically, Intel believes it’s possible to give telecom networks the intelligence to take advantage of the cyclical nature of network use and dramatically improve power efficiency. For example, its Xeon Scalable processors have features that allow customers to save energy in periods of lower demand — the reaction time of the processor means that idle cores enter a sleep mode when customer traffic is low, and resume normal operation when traffic increases. Think of this like power-saving lights in a facility; they remain in a sleep state when no one’s around but switch on as a person enters the area. The lights are powered down until needed — the lighting system is dormant but still aware thanks to its motion sensors.
This capability can be combined with artificial intelligence to predict network traffic patterns, delivering even more advanced energy-saving solutions. Ms Panhorst talked about how Intel’s application-level awareness enables intricate real-time responsiveness, sharing the example of Deutsche Telekom. The operator used servers based on Intel processors at 5G base stations, demonstrating the use of network monitoring to support dynamic reaction to changing demand. The net result was a 30% to 40% reduction in power consumption over a trial period.
Thinking back to the lights analogy, this would be more akin to the lighting system not only detecting motion but also having a good sense of who’s walking, how fast they’re moving and how long they need the lights to be on for, then adjusting accordingly. In other words, it’s a much more intelligent and intuitive system that delivers a finely tuned experience, learning from past activity in order to predict future needs.
Another example from Intel referenced a trial with Japanese telecommunications operator KDDI, which used Xeon Scalable processors in the data centre housing its 5G communication facilities. KDDI saw 20% to 40% power savings over a 24-hour period, using network monitoring thresholds to meet performance indicators while lowering energy use. All operators are becoming motivated to take advantage of these capabilities.
Although the sector’s emissions are rising, it has great potential to be a positive influence, enabling increased energy efficiency and improved environmental performance in other industries.
According to Mats Granryd, director general of the GSMA, 5G can be a major driver of sustainability in other industries, particularly when used in tandem with other emerging technologies such as edge computing, automation, artificial intelligence and data analytics. Adoption of 5G will help accelerate digital transformation in enterprises and industry sectors, supporting innovative new uses and operational efficiency . Mr Granryd’s argument is that despite rising data traffic the mobile industry will overall deliver carbon emission savings by helping other industries reach their net zero targets.
For example, more dynamic allocation of mobile resources in an organization could entail supporting more efficient supply chains, enabled by predictive analytics. It could also support automated, intelligent vehicle management in all sorts of traffic, logistics and transportation environments. Alternatively it could be used to enable data processes that reduce travel, such as remote medical diagnosis, drone-based site surveying and virtual training or tourism. A good example in the transport sector is Intel’s work with QuayChain, building a private network for the Utah Inland Port Authority. The latter has been able to reduce port congestion, minimize truck idle time and streamline cargo handling thanks to smart sensors and data analytics.
Such uses could also directly apply to environmental protection. Ms Panhorst offered an example of 5G and edge computing in rivers, where cameras are used to monitor fish, sending visual inference data to the edge data centre so that decisions can be made to reduce use of pesticides, water waste and so on.
Although 5G could be the foundation for a digital economy, that economy must also be green. The IT industry must strive to further lower energy consumption in transmitting, processing and storing rising volumes of information; to do this it needs to provide infrastructure that’s more intelligent, efficient and carbon-aware — all while delivering an ever-increasing quality of service.
Twitter
Facebook
LinkedIn
Email