6G already? Not so fast. The wireless industry hasn’t resolved two key challenges that exist in the current generation of 5G, according to Benoit Pelletier, director of edge ecosystems in VMware’s Telco Cloud business.

“One is the internet. The internet is broken. And No. 2 is 5G’s too complicated, and will 6G be a solution?” he said at this week’s 6G Symposium.

Today’s network infrastructure, largely built from the inside out with most intelligence delivered from huge centralized data centers, can’t deliver the resources and architecture required to unleash the most innovative applications, Pelletier said. “The internet was designed for almost exactly the opposite of what those applications require.”

Billions of connected devices and compute-intensive applications are pulling the internet’s center of gravity to the edge, he said, pointing to the Open Grid Alliance as a shared platform designed to meet those needs.

“Current 5G deployments are mainly targeting higher-performance bandwidth and network densification with 5G new radios and augmentation of the 4G LTE network,” Pelletier explained. “Much remains to be completed for some of the broader elements that sit at the cross-section of wireless, networking, and cloud industries.”

Killer apps for 5G and 6G are envisioned and advancing, but these apps “haven’t become very prevalent because they need the scale of infrastructure to reach the global masses, otherwise they don’t quite work,” said Sunay Tripathi, Google’s head of telco and enterprise edge.

“The killer apps already exist. I think we need to somehow get these apps out to the masses,” he said, adding that it will probably take at least a decade to adequately modernize telecom infrastructure to reach that objective. “Technology wise, we’re not there yet.”

Mahadev Satyanarayanan, a computer science professor at Carnegie Mellon University, pointed to the democratization of artificial intelligence technologies as one of the biggest hurdles to overcome in this regard. 

“We have to reduce the barrier to entry for creating this class of applications,” he said, adding that companies, researchers, and open source communities must make advanced technologies easier to assemble and deploy. 

“People don't care about the infrastructure, people care about the new capabilities they enable. And so what I see is a chicken-or-egg problem, that the implementers of novel, high-value applications are holding back because they're not sure how real this 5G is going to be,” Satyanarayanan said. “Conversely, the people who are investing in the rollout of these networks, keep looking around and asking who is really asking for this. And so, breaking this logjam is I think the single most important thing that the industry could do to help itself.”

Delivering functionality, not just connectivity, at scale remains a challenging proposition, Satyanarayanan continued. Architectural decisions made in the late ‘90s to offload computation from a mobile device to a server are precluding this today, he added. 

“Back in 1997 we didn’t have the cloud, so we didn’t need to think about the edge either. Fast forward to today, this is what the unique opportunity opened up by 5G and 6G gives you,” Satyanarayanan explained. “We see killer applications for both 5G and 6G emerging from a trifecta of very challenging properties — applications which are compute intensive, bandwidth hungry, and latency sensitive — and these can happen in both cyber-human and cyber-physical systems.”  

While early 6G research is underway, much of the promise of 5G remains unfulfilled, according to panelists at the event.

“As we talk about where we are in the state of the rollout of 5G, most 5G networks today, especially the commercial networks are still tied into a 4G core,” said Mark Reudink, VP of technology strategy at Crown Castle. “We haven’t really moved into that true 5G infrastructure with the low-latency capabilities at this stage.”