DALLAS – AT&T's portfolio of hardware and software could position the company to be an important player in the 5G ecosystem.

The service provider is already on the cutting edge of 5G development with its fixed and mobile trials coupled with its work with 5G standards. But its growing reliance on software will likely distinguish it from other operators and enable it to develop different 5G use cases.

At the recent AT&T Business Summit in Dallas, Andre Fuetsch, CTO and president of AT&T Labs, noted three fundamental shifts within AT&T that will support 5G use cases. These included the move from a hardware focus to software using virtualization and software-defined networking (SDN); the move from a centralized network architecture to a distributed architecture that the company calls its “5G real-time cloud” architecture; and moving from just observing what’s crossing the network to using machine learning and artificial intelligence (AI) to provide real insights.

Fuetsch explained that these shifts, combined with AT&T’s already established physical footprint, will allow the company to support different 5G use cases. He said those use cases are expected to rely heavily on low latency that can only be supported by multi-access edge computing (MEC).

“The cloud players have data centers in dozens of locations around the world, while we have 5,000 central offices, 65,000 cell towers, and this physical proximity,” Fuetsch said in comparing AT&T’s physical presence to those of established webscale players.

Fuetsch said AT&T might not extend computing power to all of its edge locations, but that it was likely to have “hundreds or thousands [of locations] over time.”

“5G gives us the opportunity to take advantage of the network as not just a mesh or a hub-and-spoke model, but to take advantage of its computing power for applications,” Fuetsch said.

And, with the ability to extend the processing power of the network to those edge locations, Fuetsch said AT&T will be able to offer a better economical model for delivering bits of data.

“What we are doing is giving the industry an option here to basically provide a more economical way to do their processing,” Fuetsch said.