"Our entire edge strategy is based around what our customers actually need," AT&T proclaimed in a recent blog post. The post was penned by Mo Katibeh, CMO of AT&T Business and champion of the company's edge computing initiative. Katibeh told Light Reading recently that AT&T's business customers have diverse needs, and that 5G edge compute is not a one-size-fits-all solution.
"Our customers are in fact the ones that are steering us on the use cases that they're looking for that require edge computing capabilities," Katibeh said. He said AT&T sees two varieties of 5G edge compute use cases: premises-based and network edge compute.
Premises-based is just what it sounds like: The cellular network connects to on-site servers running cloud software. AT&T calls this multi-access edge compute, or MEC. Katibeh said Chicago's Rush University Hospital was one of AT&T's first customers to use multi-access edge compute and 5G, a combination that Katibeh said can offer network latencies of less than 10 milliseconds. The technology can be used to enable remote patient care, and to leverage augmented reality and artificial intelligence for student training.
"Over time you will see us announce more hardware options that have been certified to work with our MEC solution," he said, adding that physical integration into the mobile core is fundamental to multi-access edge computing.
"The second variant is what we call network edge compute," Katibeh said. "Essentially, network edge compute allows businesses that have more distributed environments [to] use either the mobile network or the fiber network, along with a cloud instance that's very close to their businesses, to be able to empower new experiences." Network edge compute, which uses metro-based servers rather than premises-based servers, offers latencies of less than 20 milliseconds. One of AT&T's first customers for this solution is Gamecloud, which uses Microsoft Azure and the AT&T network to run its games on servers around the country, thereby delivering low-latency experiences to gamers in multiple locations.
AT&T and Microsoft first announced their edge collaboration almost a year ago, at the MWC 2019 trade show. The partnership is a cornerstone of one of the carrier's three overall business cases for 5G: enhanced mobile broadband, fixed wireless and edge computing. Katibeh said that the edge computing business case is the least understood and is one that will require more partnerships going forward.
One asset that AT&T brings to the edge computing equation is its thousands of central offices and cell sites around the country. Some of these can become mini data centers, a trend that companies like EdgeMicro and Vapor are already anticipating.
Companies in the healthcare, manufacturing, corporate training and drone detection industries are all using AT&T's 5G edge compute solutions. Some are cutting costs by using on-premises servers to centralize the processing of data that was previously processed on distributed sensors. Others, like Gamecloud, are getting closer to the customer by distributing data that was previously centralized.
Corporate training is one of the most promising use cases for 5G and edge compute, according to AT&T's Katibeh. "Edge computing can ensure that the latency of it is so low that it becomes indistinguishable from reality," he said. Deloitte University in Westlake, Texas, and Samsung's Austin, Texas, semiconductor plant are both using 5G and edge compute to train workers using remote instructors and augmented reality.
Katibeh believes that over time 5G and edge computing will enable a shift in the way people move through their career trajectories because it will help people learn and perform work remotely. "All of us are living longer, working longer, and there's a need to upskill ourselves throughout our career," he said. "This tech will absolutely give rise to a platform where people can go in and list their skill sets and then workers and companies can get together even if they are not in the same city."
— Martha DeGrasse, special to Light Reading. Follow her @mardegrasse