Are We Really at the Year of the Edge? NorCal Insiders Say Demand for Low Latency May Force the Industry’s Hand
SAN FRANCISCO, CA – As we take stock of where the industry lies today, as well as where it’s headed next, one topic is perennially on everyone’s radar – the Edge. So that’s why CAPRE convened a panel discussion at its recent Northern California Digital Infrastructure Summit titled, “The Year of the Edge: Data Center and Infrastructure Strategies for 2020” which sought to get to the bottom of a few things. Are we indeed at the year of the Edge, and if so, how will we get there?
Moderator Bruce Myatt, Data Center Business Leader at Arup kicked off the panel discussion by asking his five panelists that latter question — How will we transition? Specifically, how will regional and colocation operators make this transition to a plethora of disruptive, latency-sensitive applications and technologies work?
Arnold Magcale, Founder & CTO at Nautilus Data Technologies offered the first response, which sought to situate the discussion within a broader historical context. “I was here 6 years ago. The question I was asked, was where do you see data centers going? I said that data centers were going to get smaller, faster, and move toward the Edge,” he shared, which garnered a healthy chuckle from attendees across the room.
“Now, here we are today, and the biggest difference is that we’ve caught up with the chip sets, and it’s going to trickle down to the data centers,” he continued. “So now we’re going to see the advent of the 5Gs and the smaller data centers, coupled with the hyperscale players, and it’s going to be more dense. Compute is going to go to the edge. I see no end for that. It’s just going to keep going. Our company is focused on that, like many companies that are focused on this technology. It’s very exciting.”
Next, Arman Khalili, Chief Executive Officer at VPLS and a frequent contributor to CAPRE’s Digital Infrastructure Series, chimed in with a slightly more direct, but equally important, perspective. “When I hear the Edge, IoT, all of that, well, I started on the Edge back in 1993. I never left the Edge. There’s obviously been a lot of buzz words thrown around at the end of the day, none of this matters,” he asserted. “It’s all about application by application. Some stuff is latency-specific, some is not. Our business plan is simple. We’re trying to be, not only a jack of all trades, but a master of all.”
“I work for Goldman Sachs, providing capital, anything from debt to equity to the data center space. I spend a lot of time in the industry looking at different opportunities. When it comes to the Edge, the guys in the Southeast, like DC BLOX, will tell you that the Edge is Birmingham or Huntsville. Some people will tell you that the Edge is sitting on a tower site, like American Tower. Others will be distributed all the way back into your office,” interjected Jeff Ferry, Director at Goldman Sachs. “The Edge, I think, has a number of definitions. I of course think a lot about the Edge Cloud, which kind of takes the two biggest words of the last ten years and just combines them. But it’s kind of an oxymoron since the Cloud sets up the Edge. Right now, a lot of these opportunities are early stage, but working on existing platforms.”
At that point, Moderator Myatt asked a follow up question, with a more applied framework. “How do you see digitalization falling into the edge and what unique drivers does it bring to the Edge?” he asked.
Yuval Bachar, Principal Hardware Architect for Azure, Microsoft, offered a complex answer. “We live in an era where digitalization is no longer just nice to have. It’s a must-have. We can’t really afford to continue to operate in a room where we don’t have the ability to digitize everything. We need to do remote compute and so we need to get to the Edge. We don’t want a doctor who’s doing surgery on you five miles away having to deal with a two second delay,” he mused. ‘We all can imagine what might happen there.”
According to Bachar, we have to have digitalization for everything that we do. “People want to play VR games, well then they need VR game-level latency. And that means having stuff at the Edge. Coming back to autonomous cars, drones, or whatever is coming next, that will require some kind of guidance of what to do. Some of those elements rely on batteries that don’t have the ability to rely on AI agents running and making decisions for them and to tell them what to do,” he stressed. “They actually need someone who is close to tell them what to do. All of these elements will create benefits down the line, because it is going to enable us to progress much faster. But they will require that digitalization and they’ll require that close touch.”
“Now I’ve heard a lot today about hyperscalers coming to your neighborhood. Well, the hyperscalers are not coming to your neighborhood because they want to. They’re coming because the mega centers provide a super-efficient, much better deal. But the market requires this and the edge cloud provides cloud services close to the end point, and serving applications that today cannot be served in the public cloud. If you need to push for every camera in San Francisco back into the public cloud, you’re pushing out hundreds of megabytes per second. The network just doesn’t support that. You have to do the pre-processing first,” concluded Bachar.