The Impact of Demand Drivers for the Edge: Part II of EdgeConneX’s Origin Story
by Josh Anderson
Recently, CapRate caught up with Don MacNeil, Chief Technology Officer for EdgeConneX to learn more about the demand drivers for the Edge. This is the second part in a close-up on EdgeConneX’s origin story, continuing Part 1, published on July 24, 2017.
Continuing after a conversation about bandwidth and latency, MacNeil says that, as EdgeConnex focused on cable operators, they recognized that by bringing the center of gravity to the edge, two problems would be solved. One would be that the end user would have a better experience, “because you would offload the content from the most optimal location,” explains MacNeil.
“The second problem is a business one,” MacNeil continues. “For example, take the distance between Phoenix and Los Angeles. One of our customers realized that the capital necessary to continue to augment, to pick up movies in L.A. to bring them back to Phoenix, just in a short period of time, with one application serving a million subscribers — they were able to avoid approximately $500,000,000 in capital expenditure in a projected 10 year period. These two problems are significant.”
For Phase 1 of their data center deployment, EdgeConneX set up and deployed 23 centers in the U.S., and 26 in Europe (with three more coming online). Now, they are increasingly looking at Toronto. MacNeil says that the largest drivers and biggest users of traffic on the internet are starting to see the benefits, both from an over the top provider perspective, but also from an operator perspective. “That’s what our model is about and that’s where we see it going,” he says. As EdgeConneX looks to the future, Phase II is really about the Cloud and Cloud on-ramps.
Beyond that, network function, virtualization, and the internet of things are on the horizon. “One of the other value propositions we have, and one of the advantages of being new, is that we only build data centers where our customers ask us to. We don’t build speculatively,” says MacNeil. “For that reason, during the co-design phase with our large customers — which are large content owners, Cloud providers, and increasingly now the internet of things as it begins to emerge — we are able to ideally and optimally place those data centers to serve those future applications. So a lot of thought has been going into this.”
MacNeil wants to clarify one thing though. “I’m not here to say that the historic infrastructure of Ashburn or Palo Alto will go away, I just believe that they’ll just serve a different purpose,” he says. “Like anything that progresses over time, you will see a flattening. More importantly, putting, whether that’s access technology or compute and storage technology in the right place at the right size, I think that’s what we’re here to set up and enable.”