AT&T said it is looking at trying to turn some circuits down as users migrate from legacy technologies such as DSL to fiber
Operators are facing the challenge of having to deploy more hardware to meet the appetite of users for more data while at the same time they need to try to take out as much equipment as they can in terms of a per user, said Chris Hristov, AVP of network engineering and automation at AT&T.
“We’re almost constantly fighting the battle of having to deploy more hardware which consumes more energy, and at the same time you have to try to take out as much as you can in terms of a per-user. So we went down the road of putting radio heads to sleep. That was pretty controversial a few years ago, because from a competitive perspective, people make the argument you’re taking spectrum off-air, saving on power, but at the same time are you impacting your performance. But we found that we can do that and manage that in a way that was pretty effective,” Hristov said during a panel session at the recent RCR Wireless News’s Telco Sustainability Forum (available on demand here).
The executive also noted that AT&T and traditional legacy telcos are also always looking at trying to “groom” some circuits down as users migrate from legacy technologies such as DSL to fiber. “Grooming users onto fewer elements and then removing network elements out of the footprint are some of the things that we’ve been doing, and pretty successfully actually. We had a pretty aggressive [decommissioning] effort to do that over the last few years,” he said.
“You can very easily figure out how much is each network element consuming per user that it supports. And we have that all mapped out, and so we’re able to predict, if I shut down this particular element and move all the users onto another one, this is how much I’ll actually save every year,” the executive added.
Michael Wolfe, CTO of outdoor wireless networks at CommScope, noted that base station antennas may not be top of mind when operators start to think about reducing network power consumption. “But if you think about where some of the key power-consuming parts of the network reside, they are in that RF transmission path that lives at each site. So if you think about the losses that occur in the path between those radios and the antennas that actually transmit the signals, there is a lot that can be gained by improving the overall efficiency of those antennas,” he said. “That’s what we’ve been focused on for the past few years within CommScope, looking at how we reduce losses in the antennas.”
Wolfe also noted that a lot of the passive antennas that are used in lower-frequencies spectrum bands mainly for coverage purposes, are the ones that CommScope is currently trying to optimize for future higher spatial efficiency.
“I think infrastructure obviously is the primary consumer of power here and so a significant contributor, and there’s a lot of opportunities here , especially as carriers migrate from 4G to 5G and network functionalization to cloud-native,” said Chandresh Ruparel, senior director of the 5G/wireless core infrastructure segment at Intel.
Ruparel said that Intel has been working with carriers and other players in the ecosystem over the last years and has identified several opportunities in the application level to better utilize infrastructure capabilities as well as just optimize how the applications are being deployed.
“Another area that we found to be a low-hanging fruit here is runtime power savers. When you have the workload running, you are not requiring the server to run at peak capacity all the time because if you look at the traffic patterns over a 24-hour period, it varies. If the silicon technology has the capability for granular controls and it’s able to manage transitions with near zero latency, you can actually achieve significant power savings without compromising the key performance indicators,” the Intel executive said.
Meanwhile, Jen Hawes-Hewitt, head of strategic programs and solutions of global telecom industry at Google Cloud, said that CSP customers globally are coming to Google Cloud and its partner ecosystem to see how they can apply AIML at scale using the scalability flexibility of public cloud to help them drive energy savings. “So some of the work we’ve been doing with some of our tier one operators in EMEA has been really creating a hybrid deployment structure where you’re able to build a simulation. So doing that predictive understanding where you foresee there’s going to be opportunities for energy saving,” she said.
