Are cloud providers poised to eat wireless network operators’ lunch? Milo Medin, VP of wireless services at Google, sidestepped that question at Mobile Future Forward this week, but he did underline the lengths at which carriers are cozying up with hyperscalers.

“If the cellular network moves to a much more distributed compute, and much more distributed core, and a software managed system, well, you know, that's what cloud companies built,” he said. Building services that “enable mobile network operators to operate on cloud infrastructure, instead of a bunch of dedicated devices, that's what everybody” envisions as operators transition to 5G architecture.

Moving cloud computing and other infrastructure that drives network traffic closer to the user at the network edge will also reduce operational costs, he said. “We already see that with content delivery networks (CDNs) — 90% of consumer traffic comes from CDNs.” Moving network infrastructure to the edge “has a lot of cost advantages for both the carriers and the content providers of services already,” he added. 

Scott Petty, CTO at Vodafone UK, agreed with that assessment, describing the relationship between carriers and cloud providers as simpatico. “The opportunities for us to partner and build an ecosystem with the hyperscale cloud providers is our future to be really honest — in mobile edge compute, enabling big data applications, and the services that sit around that,” he said. 

“If I'm at an enterprise looking to build a new cloud application I'm going to build that on one of the predominant cloud services and what I want from my network operators is to be able to distribute that and give me the highest possible performance, guaranteed latency, and the services that go with that,” Petty said. 

The ability to tightly integrate the network, APIs, and services that can be enabled through hyperscale applications is critical, he said. “If an operator doesn't do that, I think you’re going to find it hard to build the kind of complex 5G applications that we’ve been talking about.”

Some of those use cases, particularly Vodafone UK’s activities with Ford on a new electric car factory, Centrica automating a smart factory for fuel, and Coventry University for virtual learning, are moving from test labs into production, according to Petty. 

“It's critically important if you want to participate in those ecosystems that you open up the network as a series of APIs and platforms that the application developers can leverage and use. No operator can ever hope to offer all of the technology capabilities that a Ford would need to automate an electric car factory or Coventry would need for virtual reality,” he said.

As such, Petty expects the notion of hybrid networks, wherein network operators deliver managed service components for patching, automation, security, and network management, to rise among enterprises that may not be comfortable with those elements. “In our markets, we think that hybrid model is more likely to be successful than a binary model of a carrier network versus a purely private network,” he said. 

That also points to the need for more streamlined deployments in enterprise networks, Medin said. “Carriers are used to getting equipment with a lot of sharp edges on it that have to be configured. Enterprises are not, so we have to come up with solutions that enable 5G networks to be deployed as easily as Wi Fi networks are. And I think there's a great opportunity to build solutions for both private and public entities to use 5G networks in that way.”

Operators’ ongoing transition to 5G and the cloud also portends a dramatic change in how enterprises will be charged for connectivity and other network services, Petty noted. “If anything good is going to come out of the COVID-19 pandemic it's probably a recognition of the idea that you can differentiate service quality,” he said. “5G is going to see us move away from a cost-per megabyte or a cost-per core model into a different set of economic models that we’ll have to adapt to over the next three to five years.”

That evolution will ride on hyperscalers’ experience and strengths in virtualization and automation, according to Medin. “Part of the opportunity of virtualization is not just reducing the cost of the compute infrastructure that will have to be distributed, but being able to use tools that folks like Google and others in the cloud industry use to automate,” he said. 

Automation is essential in 5G because typical opex and capex on a per site basis cannot be maintained at the costs operators are accustomed to on existing macro systems, Medin explained. “The level of automation that's going to be required to make any of this be affordable is going to be really important.”

A sustained, industry-wide focus on the 5G opportunity for enterprise is no accident. 5G presents entirely new use cases and systems architecture for businesses in a way that very few consumers have experienced on 5G to date. 

”The big aha is not about the sort of initial 5G rollout, where the 5G light comes on and your phone doesn't actually deliver you any more speed than before,” Medin said. “As new networks get deployed, when you're deploying hundreds of megahertz spectrum, and you now have the ability to do gigabit speeds with managed latency as an option, and the rest, that abundance is going to enable people to use their devices in all sorts of new ways.”

Until then, operators will continue to rely on increased enterprise adoption to pay for the massive investments they’ve made on 5G.