When EdgeConneX first embarked on its edge data center expansion early in this decade, the idea that the edge was the next data center frontier was a new thing. Since that time, the company’s vision has proven quite prescient and nobody doubts the importance of storage and connectivity moving closer and closer to the consumer. EdgeConneX has now built out some 40 data centers across 31 markets on three continents and is continuing to invest aggressively in its infrastructure to meet that demand. With us today to talk about what the company is seeing in the edge data center space and where it plans to go next is CEO Randy Brouckman.
TR: How has your US footprint evolved over the years?
RB: In the United States, many of our original Edge facilities have grown into now big regional campuses. Phoenix, Portland, Seattle, Denver, Atlanta, and Miami were early edge opportunities that have now grown to become large and significant regional campuses. I think that’s just a natural occurrence. We do see new edge markets emerging and some of those will grow up to be big boys and girls as well. Additionally, we have seen cloud and cloud-enabled services driving expansion at the edge. That can be as simple as cloud onramps coming into some of the edge markets to actual instances of the cloud, be they private or public or hybrid, emerging at the edge. Those clouds are often very purposed clouds, such as for AI or gaming, that require high power density, minimal latency, and tremendous bandwidth. As we went forward through 2018, we saw a lot of interesting edge-centric services and ecosystem partners emerge related to IoT, 5G, and autonomous vehicles.
TR: What kinds of growth opportunities do you see internationally?
RB: Over the course of the last year, we have seen growth in Europe with our expansion into the edge markets of Munich and Warsaw. We are seeing the same Edge promise and thesis that allowed EdgeConneX to build its original 23 Edge datacenters in 24 months and bring multiple ecosystems to bear in each center’s first days of operation. European markets that aren’t necessarily the traditional big data center markets but have healthy and expanding economies are seeing those same drivers: a first wave of content going closer to the consumers followed by a second wave bringing the cloud closer to the enterprise.
We will open our first site in South America soon, again driven by some of the same dynamics. But in the US, we almost always anchored with the largest broadband network and the content ecosystem. Here, we see both content and cloud ecosystems driving the anchoring of the new Edge facilities. In South America, we’re seeing the anchoring of data centers to be as much driven by cloud and content providers as it is by the core broadband networks.
TR: Do you think the edge will eclipse the traditional data center clusters in places like Ashburn and Silicon Valley?
RB: No. Both are going to continue to grow. You’re still going to need those major traditional peering hubs and those will always be growth markets. But we have seen tremendous growth at the edge. The content guys want to get closer to the eyeballs, the cloud guys want to get closer to the enterprises, and those enterprises want cloud on-ramps and ecosystems for performance and security. In Europe with GDPR and data sovereignty, our North American-based service provider customers can have one provider that can quickly give them capacity that is consistent in form factor that allows them to go into a new market without excess stranded capital, yet with the ease of mind to be able to scale.
TR: Is the rise of new technologies such as IoT, AI, and 5G changing your view of the optimal design of edge data centers?
RB: Frankly, I think we were skating to where the puck was going to be versus chasing around what was there. When we designed and built the very first edge data centers, we anticipated a power density that would be, at that time, four to five times higher than what you saw in traditional local enterprise-centric data centers. There was always an enterprise data center player or telco hotel in, for example, Memphis. But what was lacking was a proper Tier 3, high power density, wholesale type of solution that was optimized for content, cloud, and high-end technical customers. Gaming clouds, AI cloud nodes, and all such things are superbly well engineered. They run high power density and they run very efficiently. Traditional telco hotels can’t do that kind of stuff. Many of your earlier vintage enterprise-centric data centers are a good product but just aren’t geared up for this level of cloud technology. So, the puck arrived, or for some is now arriving, and we’re at least getting some shots on goal.
TR: You have many more nodes in North America than in Europe, how does the revenue break down and how does the infrastructure differ?
RB: Believe it or not, the company is split almost 50/50, revenue-wise. We’ve had a quite huge amount of huge growth, we just haven’t talked a lot about it. In terms of number of centers, it’s a smaller footprint in Europe than it is in the United States, and so the edge opportunities are there. The US may have been a couple of years ahead in some ways, but I would say that market is accelerating in Europe now. I think in Europe we’ve been particularly adept at handling some of the unique market challenges. The edge sites in Europe are still the same order of magnitude in size — maybe two times as big as in the US but not 10 times as big. We’ve just happened to be able to provide the full spectrum of solutions from 20 KW up to 100 MW. These things ebb and flow. Last year it was a bunch in Europe, this year may be a bunch in America.
TR: Do you foresee demand coming from further out on the edge? If so, how might that affect your approach?
RB: Before we built the first edge data center, EdgeConneX had already deployed a couple hundred ‘Edge PoPs”, small micro facilities in the basements of office buildings and at cell sites. They have been and are still a profitable product line for us, and we have been investing a bit in terms of what the next generation of those microsites look like and where they will really end up. I think the original thinking that there’ll be a data center in every cell site, I think that mentality has shifted. I think we stand ready for the micro edge when the time comes.
TR: What do you see as the biggest challenge in building out the next generation of edge infrastructure?
RB: Scaling. I think one of the things we’ve been very adamant about is if you’re going to go to the edge, you have to be able to automate all your operations. You can’t staff a cell site, it’s an unscalable business. You need to be able to have a consistent set of operations: alarming, management, security, methods and procedures for both reactive and proactive maintenance, etc., and you have to be able to scale that from 20KW to 100MW. You need to be able to allow your customers to have a view into all of their deployed gear anywhere, at any site, at any time. Without that, you have a real challenge to make the edge work.
Another challenge is that you must get balanced input from the cloud providers, the content providers, and the networks as to where the next 10 most important places are that the edge needs to move to. One of the things about the edge that’s different than your traditional, big data center campus is that you have to build the data center where the customer needs the data center as opposed to where it’s easiest to build the data center. With the edge, you’re optimizing your customer’s customers experience. You have to bring the data center to them, as opposed to dragging everybody to a wonderful plot of land somewhere. That’s a different design, build, and delivery paradigm than your traditional gigantic campuses.
TR: So far, your buildout has been mostly organic, but do you view M&A as a path toward additional expansion?
RB: I’d say 98% of everything we’ve done has been not via acquisition. But for the right opportunity, we’ll buy, upgrade, and expand. Typically, when we’ve bought something, we’ve bought it because it was the right location and it had a bunch of empty space that we could properly reconfigure to a proper edge data center. Maybe it already had the networks that we needed. Maybe it already had the power and we would be able to accelerate our delivery to market. But typically, we build versus buy.
TR: What do you think is ahead for edge data center design? Are new technologies and services rising or is it just about scaling space and power?
RB: The good news is it seems to be more of the same, but keep in mind we got that first part right with high-power-density wholesale economics. Even at the edge, EdgeConneX runs a wholesale business model with its customers, which is very different than a retail data center at the edge. And that’s not a surprise if you think about the types of companies we’re talking about: the world’s five largest companies by market cap plus everything that hangs around their ecosystems. Those companies are looking for is that wholesale experience. They don’t need hand holding. It’s their gear packaged the way they want it, deployed where they want it, consuming the power they need, when they need so that they can plan their business 5, 10, 15 years forward. That’s what we’re able to deliver at the edge.
TR: Thank you for talking with Telecom Ramblings!
If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!Categories: Datacenter · Industry Spotlight