How Are Traditional Multi-Tenant Data Centers Evolving to Meet Hyperscale Cloud Growth?

LEESBURG, VA – CAPRE’s seventh annual Mid-Atlantic Data Center Summit sought to shed some light on the impact of hyperscalers on the colocation game, so we gathered a crew of highly respected thought-leaders who are active in Northern Virginia to opine on how hyperscalers are changing their business model. The panel How Traditional Multi-Tenant Data Center Operators are Evolving to Meet Hyperscale Cloud Growth: Challenges & Opportunities sought to showcase both the challenges that the hyperscalers have brought, and the opportunities.

Moderator Robert Dennelly, Director of Engineering & Product Line Management at AFL Hyperscale kicked off the panel with some first-hand insight. “We’ve deployed a lot of fiber optic cables at a lot of these core data center sites for hyperscale customers around the world, but more and more, we’re seeing hyperscalers utilize colocation facilities,” he shared.

“One of the things I was really curious about is that, in our experience with hyperscalers, they’re doing data center projects at a scale that is so much different than in the enterprise space,” he continued, listing for example, the amount of fiber that they’re deploying and the data rate speeds they’re approaching. “We have a lot of discussions with hyperscalers where they’re already onto 100 Gig, and now they’re approaching us about 400 Gig. So there are challenges and opportunities for our business, and I suspect the same is happening in the colo market.”

“So my first question for the panel is, regarding the profile of these hyperscalers,” he ventured, specifically seeking out the expertise of panelist Steve Conner, Vice-President for Sales & Solutions Engineering at Vantage Data Centers. “Is there a model or an infrastructure fit that’s a standard design? Or do you have to redesign your colocation facility to something different to accommodate them?”

Robert Dennelly, Director of Product Line Management & Engineering, AFL Hyperscale

“That’s a great question. What we’re seeing is that it’s a little bit of both,” replied Conner. “When we look at data centers in general, even the core buildings, these guys are buying at the size of 6, 10, 12, up to 36 or higher megawatts. Where we’re having to make changes is on some of the infrastructure inside of the building and also on some of the infrastructure outside of the building.”

“When it comes to connectivity, I can’t tell you how many times I’ve been asked, how far away are you from these guys or those guys [referring to his co-panelists], and we’re starting to have to rethink the way we’re [building] data centers,” he revealed. “Five years ago, you just plugged some fiber in, and there you go. Today, I have to start thinking about fiber distance from Point A to Point B, I have to start thinking about latency, and how we’re going to deploy based on them. and those are just a couple of examples.”

In other words, according to Conner, the bottom line is that the core building itself of a data center isn’t really changing as rapidly as the infrastructure on the inside (and outside) of them. At that point, Conner’s co-panelist, James Leach, Corporate Vice President for RagingWire Data Centers, offered his point of view to list another key change.

“I think you’re spot on,” he concurred, before diving deep. “I think we’re having to respond to all of these things with hyperscalers, but one thing we’re experiencing is that you have to think about how you secure your facility. And a number of the hyperscale folks have specific security requirements, not only from the technology side of things, but from the operations [side].”

“The security staff that you have might have to be augmented,” Leach suggested, preparing to efficiently conclude his remarks. “The processes that they follow might need to be changed. And the way the security is delivered might be extended beyond the facility actually to the equipment that’s out on the floor.”

E-mail me your stories,  industry news tips, and press releases.