Generative AI and Automation Seen as Unavoidable for 5G Networks

Companies are still figuring out how they’ll use generative AI and how it will benefit them as they move to greater automation, Rode Kirk, Microsoft global sales director-media & communications, Americas, said during a Fierce Network webinar Monday. Other speakers said as networks become more complex, companies will have no choice but to embrace automation.

The key question is which “partner” a company selects as its AI provider and which frameworks it has that can be used “out of the box,” Kirk said. “You have to be careful" about "whom you select for your AI. How that AI is developed and contained within your organization, with your data boundaries, and who has access.” Companies should “take a step back” and ask “who we are as an organization" and "how do we operate? What direction do we want to go in?”

Companies must adopt a growth “mindset” and be willing to embrace change even though front-line workers are worried about losing their jobs to AI, Kirk said. He said, though, that companies see AI more as a way of gaining a “competitive edge” than cutting costs. “We need to embrace this, just like dotcom in the '90s.” Kirk estimated it will take providers a decade to figure out what AI can do for them.

The changes in wireless networks as 5G expands over the networks of the past are driving the need to automate, said Nandan Atur, technical director-cloud at Rakuten Symphony, an open networks company.

Atur discussed how networks have evolved from 1G in 1983 to 5G today. With 5G we have ultra-high definition content on the network and massive machine-to-machine communications, which open "possibilities for industry 4.0, smart grids, smart cities and maybe even autonomous driving in the next few years," he said. Industry 4.0 is a widely used term for smart manufacturing.

“The type of communication and the depth of communication has significantly progressed in the last 40 years,” Atur said. There has been “a paradigm shift toward being cloud native.”

Atur argued that automation is critical. Managing thousands of network configurations “is a big challenge.” With millions of elements in the network, carriers need a “granular view” of what’s happening with each one. When a network has billions of devices and users, “updates are super important” and the network can’t be disrupted at any point.

For Norwegian provider Telenor, moving to a “zero-touch” automated network is tied to the goal of offering “everything as a service,” said Pal Gronsund, director-cloud strategy & architecture. “There needs to be a lot of automation within each layer” of the network.

The biggest challenge is getting people onboard and changing corporate mindsets, Gronsund said. Providers must put “automation first, everything as a service first in whatever you do,” he said. They need to build a “culture embracing this change and way of work.” Carriers need to be able to tell a system how the network should behave, using some kind of application programmable interface, and probably a standard data model, “then the system should take care of the rest, like scaling, optimizing, healing.”

Carriers face technical and cultural challenges when introducing automation, said Ram Ramanathan, senior director-product, at tech provider Ribbon Communications. Any telco “should be aligned, top-to-bottom, left-to-right” if it wants to move to automation, he said: “Without alignment, it’s going to be very difficult to achieve some of the low-hanging fruit and then build from there.” Automation will happen in steps, he said.

“People are part of this, even though we call it automation,” Ramanathan noted. The skill of employees “may need to be enhanced” and steps to transformation “need to be laid out,” he said. “The technology and the people go together, and the expectations need to be set the right way.”

Panelists also discussed the importance of open-source systems like Kubernetes to automation.

Under an open-source system, everyone can access the source code and “the community and the ecosystem around it, so you can talk to people,” said Ildiko Vancsa, director-community at the Open Infrastructure Foundation. The knowledge sharing part “is really huge, and I think a lot of people overlook that,” she said. “You have access to experts all around the globe, and they don’t have to work for the company that you work for.” It also helps people “understand much better the fundamentals of cloud computing.”