Open RAN opens the door for network automation and service differentiation
Network and software providers would find it more useful to have less choice at every stage of future RAN specifications as the wide menu of options usually represents a challenge for companies, Picocom Chief Solutions Architect Doug Pulley said at the recent RCR Live: Telco Reinvention event in London.
“Over the years, 3GPP has become increasingly a menu of choices rather than a solution ready to implement” he said. “And that’s great because it sort of builds in potentially some flexibility, but from the point of view of somebody building the components and the software that need to go into equipment and then, bearing in mind the test and interoperability required, things like Open RAN, this just makes everybody’s life a lot harder.”
Pulley continued: “You know, choice is the enemy of decision. I think it would be much more useful perhaps going forward to have less choice at every stage of RAN specification, but to be able to substantially revisit the specification on a more regular basis than once every 10 years, or once every five years.”
Commenting on the role of artificial intelligence (AI) in Open RAN architectures, Pulley noted that there are currently two areas of focus. “There’s AI in the network and then there’s AI for services and traffic. When we look at AI for the network, AI needs to exist in a number of forms. For automation purposes, AI needs to be scoping the network to hoover up data or train models and redeploy them. There’s a lot of discussion about using AI to build pieces of the RAN itself to actually accelerate or provide new functionality to make the RAN just work better.”
Pulley also said that the opportunity for AI in the Open RAN field is that the disaggregated architecture provides a good platform to try out AI and machine learning (ML) pieces. “AI and machine learning is comprehended by things like the RAN Intelligent Controller (RIC) from the outset, so it provides a platform to be able to rapidly deploy solutions from maybe smaller players so that they can be field trialed and proven and trusted much more quickly than perhaps waiting for more monolithic solutions from the usual incumbents.”
A key focus area for Open RAN specifically, 5G in general, and 6G in the future is leveraging automated decision-making to reduce network energy consumption. Pulley pointed out that discussions around 6G are looking at an energy per bit reduction by a factor of 10 while there’s also an expected 10x increase in network traffic. “So there’s no net gain here. I think there’s a definite rethink required and there’s definitely opportunities for generative AI to massively reduce the amount of information that actually needs to be sent over the network.”
