Carriers worldwide, from NTT Docomo, Telefonica, and Vodafone to T-Mobile, Verizon and AT&T have committed to being carbon neutral or to reduce their greenhouse gas emissions to Net Zero by 2030 for some carriers, and by 2040 for others. With these carriers committing to aggressive timelines, major changes will have to be made to the GHG emissions of RANs, according to a new report from Mobile Experts. From simply shutting down 2G networks to deploying Artificial Intelligence/Machine Learning (AI/ML), carriers have a lot of tools to cut electricity use.
The progress of making RANs more efficient is a series of good news/bad news scenarios. One piece of good news is the industry has already reduced a lot of its electricity usage in the transition from 2G, which requires power-hungry HVAC units, to LTE networks, whose remote radio heads are passively cooled. The bad news, reported by Mobile Experts, is the efficiency improvements in semiconductors and power amplifiers will have less impact in the future.
“Continuing the improvements of the past 30 years, we can expect some ongoing efficiency improvements from smaller CMOS process nodes and from Power Amplifier techniques, but we believe that these will have less impact in the next 20 years,” the report reads.
Wider spectrum bands and more efficient coding/modulation have also been responsible for significant gains in efficiency in the past, but that too faces diminishing returns in the future, according to Mobile Experts.
“Experts all over the mobile industry tell us that 5G is getting pretty close to Shannon’s Limit, meaning that any change to a 6G coding/multiple access scheme may only squeeze another 5-10 percent of additional data from each MHz of spectrum,” the report reads.
But there is hope for increased RAN efficiency in the prospect of improvements made possible by AI/ML, self-optimizing networks (SON) and Open RAN Software. SON software has already been used to save energy by turning off radios when not in use. AI/ML may take that many steps further, allowing the RAN to turn off cells, change frequencies or even switch modulation techniques in real time to operate on the least energy possible. Optimization may be made on a per-user basis, not just on a per-cell site basis, the report notes.
Perhaps the most hope for efficiency in the future comes from the development of Open RAN networks with near-real time RAN Intelligent Controller software, which brings multi-vendor interoperability, intelligence, agility, and programmability to the RAN. Open RAN will bring together “hundreds of smart engineers” to work on the issue of energy consumption, according to Mobile Experts.
“We believe that this will, in the end, save even more power than the pet algorithms developed in a major vendor’s lab, because no single company can think of all possible ways for an xApp to optimize power consumption for the nearly infinite variety of users,” the report reads.
By J. Sharpe Smith, Inside Towers Technology Editor
Reader Interactions