Everything is moving closer to the edge, but not for the sake of adopting edge computing technology. The advent of the edge is about delivering better products and services more efficiently while maintaining data sovereignty and security, according to Red Hat Chief Strategist Ian Hood.

Those goals serve as major drivers for telcos in particular to take "advantage of partnerships in the cloud that might have edge solutions that suit this picture," Hood said during an STL Partners webinar.

Telecom operators need various types of applications – just like any other business. These include business apps and services, customer interfaces, operational and billing systems, media networks, and 5G core networks that extend across wide areas. "The question for the operator is, as I look at this real estate, how do I actually decide what percentage of that stays on premise for what purpose, and what percentage lives in one or more of the public clouds?"

Hood explained that in most cases where telcos have a certain set of requirements, they pick one or more hyperscalers like Amazon Web Services (AWS), Google Cloud, or Microsoft to deliver those services, or "potentially deliver it in partnership with them over the edge and deliver that extended distributed picture processor secure network," Hood explained. "This is really the direction that we're headed."

The economics of partnering with a hyperscaler for edge services should be a significant aspect of consideration, he added. Telcos should look at what the cloud provider's costs are in contrast with how fast "I can go spin it up to do my first thing."

In addition to telcos, essentially every industry is active in partnering with hyperscalers for edge computing. In healthcare, for example, edge computing is used to communication between facilities and perform brain scans, lung scans, and other diagnostic surveillance services. "Another one that everyone's all excited about is autonomous driving," he added.

"These customers have a particular set of requirements from a data and application perspective that might fit one or more of the hyperscalers. The reality is that we already have in almost every operator at least two, if not three, hyperscalers in play," Hood said.