AT&T’s recent push to extend its edge zones to a dozen markets by year end highlighted the ongoing complexity operators are having to manage in rolling out their 5G networks.

Those edge zones are simple in design: they are AT&T servers living in a connected location near a hyperscaler data center and that can be managed and controlled from a centralized resource. This allows those zones to be physically closer to end users to support low-latency use cases and can share compute and storage resources as needed with that hyperscale infrastructure.

But how those then play into AT&T’s broader 5G network deployment plan is where it gets more complicated.

The ultimate goal for operators – at least today – is to run a single 5G network core that can be managed and controlled to support all of the customer and traffic needs of their network. This would include housing all of a carrier’s software, hardware, and spectrum resources.

The reality is that most networks are running a mix of network resources spread across different back-end and technology stacks. This can include 2G, 3G, 4G LTE, and different 5G iterations based on standalone (SA) and non-standalone (NSA) network cores, not to mention various radio access network (RAN) vendors.

Throw in an expansive physical infrastructure provided by a growing list of vendors and the challenge grows.

AT&T has some of the longest history in this having kicked off its disaggregated and distributed network efforts with its Domain 2.0 program in late 2013.

One of the initial tenants of that program was to “increase the number of suppliers and partners AT&T can do business with.” This was important as it provides the industry with a larger number of competing vendors for resiliency and sourcing and also provides more competition to help expedite innovation.

Igal Elbaz, SVP and network CTO at AT&T, explained in an interview with SDxCentral that this history has helped the carrier in this move.

“The biggest transition is that you are moving into a complete end-to-end software and cloud-native architecture,” Elbaz said. “For us that transition was less of a challenge because we have been in this virtualization and SDN journey for almost a decade. So the transition of physical boxes, fully integrated into software running on a network cloud infrastructure is something that we're fairly familiar with.”

Despite that familiarity, there have also been challenges, “but it’s an exciting challenge,” Elbaz added. Much of that excitement has been in getting these pieces from different vendors to work as one.

“You need to have a certain amount of maturity across the ecosystem in terms of carrier aggregation on specific spectrum position and that has to do with the RAN software and the chipset and the devices,” Elbaz explained. “So all of these need to come together in order to make sure that when we have the opportunity to move a customer to a standalone infrastructure, they have at least an equal experience to what they are used to, if not better.”

Despite these vendors, partners, and locations, Elbaz reiterated that “AT&T’s 5G mobility core runs on AT&T hardware in the AT&T data centers and operated by AT&T.”

That’s not to say that AT&T has not looked for help further up the stack. The carrier famously moved in this direction in 2019 when it struck a deal with Microsoft to support the carrier’s non-network applications.

This eventually expanded two years later in AT&T selling its internally developed Network Cloud technology to Microsoft. This now has AT&T running many of its 5G-related services on Microsoft Azure.

Elbaz said this deal was predicated on AT&T taking advantage of the cloud ecosystem.

“At the heart of this is how can we take advantage of the hyperscaler investment in tools, in automation, and cloud capabilities, because that’s their business and they are really good at it,” Elbaz noted. “And knowing that this is how we are going to run the network going forward so why not take advantage of this.”

AT&T is looking to further take advantage of that work as Microsoft further builds out its Azure Operator Distributed Services (AODS) platform that at its heart is AT&T’s Network Cloud.

“That combined thing is what we call the carrier-grade AODS stack that we are going to start implementing once they are ready,” Elbaz said.

Microsoft’s efforts echo those of its hyperscale and cloud rivals that have increasingly inched toward the telecom space. While none have quite gone full operator, they are getting close.

The biggest move might have been Amazon Web Services (AWS) recent commercial launch of its Private 5G platform. That service, which has been gestating for nearly a year, integrates small cell radio units, Outposts servers, a 5G core, and RAN software running on AWS-managed hardware.

Elbaz said these efforts are good for the broader ecosystem, but he does not think hyperscalers want to make the full jump to being a network operator.

“I don’t think they want to be operators, they just want operators to use their services,” Elbaz said.