How are telcos embracing AI in the RAN?

By Catherine Sbeglia Nin September 13, 2024

Collected at: https://www.rcrwireless.com/20240913/fundamentals/how-are-telcos-embracing-ai-in-the-ran

According to Blue Plant’s VP, telcos will lean into AI in the RAN much more ‘aggressively’ in the 6G era

Even as 5G has yet to deliver on many of its promises, the balance is shifting; more and more conversations are beginning to favor the topic of 6G and how AI will be a cornerstone of this next generation. And it’s true that embracing AI will be key to achieving the RAN modernization necessary for 6G systems, but what are telcos prioritizing as they introduce more and more AI capabilities into the RAN?

An open, democratized AI environment

According to Blue Plant’s Vice President Kailem Anderson, telcos will lean into AI in the RAN much more “aggressively” in the 6G era. “In 6G, you’re going to see the vendors and more importantly the telcos really embrace the openness in RAN in terms of disaggregation of the equipment — it’s not there yet in 5G, let’s be honest,” he said, adding that as the industry moves from 5G to 6G, there will be “true disaggregation of the RAN.”

Further, telcos today rely on the AI capabilities built into the software they purchase from their RAN vendors, which results in a closed environment and a lack of control for the telco. “I believe, moving forward, you’re going to see this democratizing of AI play out not only in the RAN but in the entire network where instead of being beholden to the vendors that create the algorithms, the AI will open up,” Anderson told RCR Wireless News.

In this scenario, telcos will be able to bring their own AI and apply it to the data sets that are being leveraged within the RAN, which he predicted will result in a new dynamic, leading to the rapid acceleration of AI within the RAN.“Being able to leverage that data for customer care systems, or to drive value-added automation, maybe even automation that is inter-carrier — from carrier-to-carrier — because you’re not innovating just at the pace that the vendor is providing their AI into their equipment,” he said, “but you can accelerate on top of that.”

Reducing energy consumption

Another priority for telcos as they consider the future of networks, according to Alex Jinsung Choi, chair of the AI-RAN Alliance and principal fellow of SoftBank Corp.’s Research Institute of Advanced Technology, is the reduction of OPEX.

“And energy consumption is one of the biggest factors of OPEX,” he continued. “AI will play a big role in making RAN operations more sustainable. Through predictive analytics, AI can predict the most energy-efficient times and places to operate network components; it can switch systems into low-power mode during times of low usages, such as overnight to conserve energy.”

But, according to David Soldani, the SVP of next generation advanced research at Rakuten Mobile, telcos need a greater degree of operational freedom in order to realize energy savings. Currently, the only real option for power reduction involves switching a monolithic base station or cell site on or off. “But if you have really a composable, disaggregated infrastructure, where there is a decoupling between the hardware and software, but more importantly the platform and the workloads, I now have a high degree of freedom and can achieve optimal energy consumption at the given rate of service for the end user,” he told RCR Wireless News.

He explained further that switching cell sites on and off invites a “tremendous” amount of risk, and so, telcos want AI to deliver more “liquidity” when it comes to energy consumption by using it to predict — at various network levels — where such an action should be taken. “For example, [Rakuten is] using AI to determine if certain CPUs in the hardware can be switched off … that’s where AI contributes, down to the hardware. And if the platform supports it, we have ability to go a level up to scale horizontally or vertically the resources you provide to your nodes … You can measure your energy and then move your workloads to provide less or more resources, vertically to the nodes or horizontally to the cluster, so that you have an optimal way of consuming energy,” said Soldani.

Finding a ‘means and method’ to revenue

Soldani also laid out a few key goals for 6G, but highlighted that one of the most important for Rakuten — and presumably, for all operators — is establishing a “means and method” to revenue.

In service of this goal, AI-RAN Alliance’s AI-and-RAN working group is exploring how using the same infrastructure to run both RAN workload and also AI workload simultaneously will open up new revenue streams for telcos. “This is about making the most of what we have … to support AI applications while still managing our networks core functions,” said Choi. “The outcome from this group is to show us how to increase resource utilization and open up new revenue streams by hosting various AI applications on the same platforms that runs our network functions.”

Continuing to build trust in AI

Anderson is confident that in the future, AI will be “pervasive in all aspects of the network,” but added that right now, telco’s lack of trust in the technology remains an obstacle. “It’s fair to say that there is a heavy level of skepticism on the part of telcos around AI and explicitly, operations teams don’t trust it yet,” he said. What is AI automates off a false positive or off a false negative? What is this blows up my network? These are the questions telcos are still asking today.

Therefore, operations teams only implementing what Anderson called “partial-closed loop AI” use cases, where an operations team will get the output and confirm it implementing the change. AT&T’s “co-pilot” approach to AI comes to mind. The telco uses AI as an additional network site quality agent and to provide its ops teams with recommendations around what, if any, network parameters should be changed.

“I do believe that in the 6G world, AI will be embraced with a trust that isn’t there today. I think the biggest shift you’re going to see is not a technology shift; it’s a people and trust shift as it relates to the adoption of AI in a 6G world. And I truly believe that between now and 6G, this trust is going to evolve to the point that ops teams will start to give up that control,” said Anderson