- MEC may have flopped in the 2010s, but Ericsson argued the technology landscape and demand drivers have changed
- AI traffic growth is changing the edge equation, with surging mobile and uplink data demands expected to make centralized cloud routing slower and less economical
- Ericsson argues modern networks can now support the low-latency, high-reliability use cases edge computing was built for
The mobile edge computing (MEC) efforts of the 2010s were a notorious flop for operators. But the sands have shifted, and edge technology could soon go from ahead of its time to showtime.
We’ve written before about how the AI edge opportunity for operators is complicated. And leading voices in the industry – including TM Forum’s CEO – have told Fierce operators shouldn’t bet the farm on it. But Ericsson Americas Chief Strategy and Technology Officer Joe Constantine has a different view.
MEC, he told Fierce, was early but it wasn’t wrong. And a lot has changed in recent years.
“Ten years ago, MEC was a supply side concept,” he said. Now though? “We have AI inferencing. It's the growth opportunity, an application that MEC did not have at the time. So, if you look at this, we believe that the industry, the technology and the market is vastly different today from 10 years ago.”
How have edge networks changed?
AI isn’t just a use case, it’s also a driver of massive amounts of data traffic, Constantine said. He noted that between 2023 and 2029, global mobile traffic is expected to triple, per Ericsson’s Mobility Report, with AI the key driver of that growth. Ericsson has also predicted that uplink traffic in particular will grow by 10x by 2035, as AI shifts the dynamics of how information is shared back and forth across networks.
“Routing all this traffic to a centralized cloud isn’t just only slow, it’s economically not even sustainable,” Constantine argued.
Networks have also changed, he continued. When the first iterations of MEC hit the scene, networks were built for “best effort” connectivity. But today’s 5G networks are built for time-critical communication and are capable of supporting 15 millisecond latency and offering five nines of reliability.
“That’s a vastly different network capability than we had before,” he said. Couple that with 5G network maturity and the rise of technologies like 5G standalone, network slicing, cloud native cores and new tooling, and suddenly the MEC equation looks a little more solid, he argued.
What use cases require MEC?
Pressed about what use cases really require MEC, Constantine pointed to physical AI and robotics systems as examples.
“Any system that moves – so drones, vehicles, humanoids, robots talking to other robots…all these will become an autonomous system that needs a network to think with it,” he said. Put another way, it won’t just be about delivering compute at the edge but the broader contextual awareness that networks will be able to offer.
Constantine used the example of an autonomous car operating with an onboard LIDAR system. The car can see what’s in front of it, but not what’s around the corner. That’s where future network sensing capabilities at the edge will come into play.
“I don't think the device itself, without any connection to other devices and without any connect to the real-world models and context awareness, can actually operate in a society,” he concluded.
Read more about the edge debate here:
Opinion: The telco edge opportunity is real(ly complicated)
TM Forum's CEO says edge is the 'billion dollar question' for telcos