Distributing compute closer to the end user opens up the latency benefits of 5G
The conversation about the relationship between 5G networks and edge computing has been going on for some time. The premise is relatively straightforward: To take advantage of the latency benefits that will come up with next-generation cellular, functions generally associated with centralized data centers need to be distributed in support of latency-sensitive use cases like mobile VR, autonomous driving and so forth.
And although the premise is straightforward, the nuts and bolts decidedly are not. Earlier this month at the Consumer Electronics Show in Las Vegas, Nevada, I explored the relationship between 5G and edge during a panel discussion hosted by Intel. Here we’ll recap that conversation and hopefully shed some light on what needs to happen to light up the edge and ensure 5G lives up to the hype.
To hit some of the high points of the panel discussion, read on; to watch the whole thing, here you go. To get the complete picture, do both.
Where is the edge?
Intel’s Caroline Chan, vice president of the Data Center Group and general manage of the 5G Infrastructure Division, part of the Network Platforms Group, explains: “When we looked at edge, we looked at where it actually touches…theuser experience, the delivery. It could be on-prem, so that could be millions of…access points on-prem, or it could even be off-prem. Mainly [when] we look at the edge, it’s based on latency–what sort of latency do you need to deliver the services? 5G with edge compute is the biggest payoff of digital transformation.”
Who controls the edge?
Asked if operators would own their own edge infrastructure, buy or lease it following an as-a-service model potentially from new market entrants, or look to established cloud service providers, Iain Gillott, founder of president of research firm iGR, said: “All of the above. It really depends on…where that edge is and the latency. Some applications are going to be very, very low latency–autonomous vehicles…robot control, things like this, and you’re going to want to get as close to the radio as you can. In that case you’re going to work with the operator very, very closely because they’ve got the radio. In other cases maybe you can be a little further back in the network at a little bit of a bigger cloud data center maybe at a city level or a neighborhood level. Maybe now it’s outside the carrier network but you connect in. Yes is the answer, but it’s going to vary by application. I think it’s going to be difficult in some ways for the cloud providers to say now we’re going to be in the radio business and get all these sites. And likewise it’s very difficult for the carriers to say we’re going to be in the cloud business and go build massive data centers. They both need each other. There’s a natural partnership there.”
What’s the relationship between edge and cloud?
Rob Topol, general manager of 5G Advanced Technology in Intel’s Technologies, Systems, Architecture and Client Group, said: “With 5G you’re seeing essentially a technology that was born after the cloud. It’s the first wireless technology that can take advantage of the orchestration structure that’s been in place for cloud and servicing users. And now taking the virtualization technology that exists in a data center of how capacity is allocated and you’re moving that down in this virtualized capacity…at the edge of the network. And now bringing that as a way to bring extremely low latency use cases, applications and services that 5G essentially brings. And that’s of course in partnership with the enhanced mobile broadband aspects of 5G. When you combine those two, now with the fact you’ve got an edge computing platform nearby, the new applications and services are really limitless.”