Compass EdgePoint’s Sharif Fotouh Talks Evolution of Provider Side, Impact of Edge Computing in Advance of CAPRE’s Denver Summit
by Josh Anderson
DENVER, CO — Sharif Fotouh, an ex Googler, and founder of EdgePoint Systems, now Compass EdgePoint, is a recognized voice across both the information technology and the data center industries. Sharif has more than 10 years of tenure leading large data center and technology teams, and has leveraged that experience to launch Compass EdgePoint. At Google, Sharif founded and led Google Fiber’s national network facilities and deployment engineering program. In previous roles, Sharif has led both data center operations teams as well as global IT implementation projects. Fotouh will be a featured speaker at CapRE’s upcoming Denver Data Center Summit. In anticipation of the conference, we connected with Sharif to talk about his favorite topic — the Edge.
CapRE: Thanks for chatting with us, Sharif. What’s been keeping you and Edgepoint busy lately?
Fotouh: Compass Edgepoint is part of Compass Edge Solutions (CES). CES is focused on providing a solution that is one step beyond, really many steps beyond, the box or the container. I think that the edge space is dominated by players focusing on that container and consider that the entire solution. But it’s more than just the box, people have been building modular structures for telecom for decades. There’s definitely innovation there. but when you look at the challenge, or looking back at my experience with Google, I like to speak of that Oh crap moment – when you call up someone who can make you 500 containers a year, and they reply, “Here’s an empty Excel spreadsheet, just fill out the different addresses you want them shipped to.” And then you realize you have 500 different construction projects! That’s a beast.
The problem that we solve for at CES is beyond just the shelter itself. The challenge is how to support the whole lifecycle, including maintenance to monitoring. We want to provide that holistic solution. It ties into a picture where we can sell you 1 or 2 megawatts, but that megawatt comes in the fashion of 100 shelters across the country, and it’ll be just as easy for you to contract that as end-user as it would be to go to get 1 megawatt from a wholesale provider. Being able to say yes to that order and making it as turn-key as possible is no small feat. Fortunately, we have enough experience to do it, the scars on our back, you could say, and the lessons learned.
CapRE: Thank you. So let’s talk about the Edge. Where’s it going next?
Fotouh: The Edge has bubbled up and started to get really frothy. Alternatively, we like to talk about in pragmatic terms — teasing it away from Virtual Realty, Augmented Reality, and the autonomous cars conversation. I’m excited about those too, but they aren’t coming around for widespread use in the next year or two. So trying to push the conversation into more immediate applications has been pretty important and is a key focus of ours. These discussions can go sideways when someone brings up a buzzword like VR. Not to say I’m not excited about VR going mainstream since you can consider me someone who has way more Virtual Reality toys than I care to admit. But honestly, those applications are still early-stage and we don’t know the infrastructure needs of those applications, yet if and when they becomes available for mass use.
CapRE: And what about how the Edge has changed?
Fotouh: The Edge is not a new concept. There has always been some level of compute that has needed to be in a specific geographic location with an orientation to the users that the application is serving. Ten years ago “the Edge” was putting in an East Coast and West Coast footprint, maybe one in Europe. An application provider would think, that’ll take care of most of my customers. That was the Edge of their network, and it was pretty valid. As we’ve shifted across that spectrum, over the past few years we are seeing footprints go beyond the East Coast and West Coast, we’re looking non-traditional data center markets, local caching and peering, etc.
And as you start to plan for that Edge, and look at markets outside those considered “Tier One” you realize the capacity and availability for decent data centers in those markets drops off pretty quickly after twenty or thirties cities in the US. I experienced that first hand at google when looking for DCs in Louisville or Oklahoma City facility, couldn’t find a decent N+1 facility that could sell me 20 racks. Now we’re looking at where the Edge goes next, and that spectrum continues.
The penetration of these large networks and these large CDNs and global apps has to reach more cities than those top 20. I think there’s a prime argument for getting into Tier III cities – that’s long-tail to be chasing. This is just an economic argument – not so much that everyone just needs ultra-low latency.
CapRE: What else have you been thinking about lately?
Fotouh: With enterprise users, there has been a focus on “cloudification.” And that’s great because they can take advantage of scaling instantly with an AWS sign-up page. It’s amazingly enabling and freeing for organizations to focus on something other than racking servers and cabling up, but the challenge with clouds is that you see stranded remnants of their footprint that doesn’t belong in the Cloud.
CapRE: Can you share some examples?
Fotouh: Look at a healthcare facility, for example. Most hospitals have pretty sizeable footprints – 50-100 racks. They’re built up pretty decent sized room to support everything they needed. Well what happened is most of their footprints have migrated into the cloud but there are still 15-20 nagging racks in those rooms that you can’t move to the cloud. You need local records in case there’s an outage of natural disaster. Regulations dictate that they need to be cached locally to continue to provide services. Computer controlled surgeries are very exciting, but they need to be performed locally. They can’t be in the Cloud.
The conversation we’ve had with them has been about putting a shelter on premises, all of the 2N hardened capacity that you’d expect it to have. It won’t be flimsy – it’s wind-rated, 2N throughout. But now it’s taking 10 parking spots and you can free up that 100 rack datacenter room for more patient beds. SImilarly, when you talk to the industrial manufacturing sectors, a lot of our clients aren’t in areas with big data center capacity. And I hate to say “Industrial IoT”, but putting sensors and robots in a manufacturing facility means their data will explode in size.
They could carve out a room in their facility which would take valuable real estate of their assembly line, or they could just put something out in their parking lot.
CapRE: What are you most excited about at our upcoming Denver Data Center Summit?
Fotouh: I’m excited to share our perspective on where the focus is for the Edge in the near-term. I’m excited to hear what the industry is looking at and what perspectives my colleagues have on the edge, and the data center landscape in general. Every time I’m at a CapRE event, I realize that there is such a broad spectrum of industry experts in the same room, which breeds a really interesting conversation.
5G has drivers and dependencies for Edge computing and I’m definitely interested to hearing about a 5G plan and what it needs from a compute and infrastructure perspective, as well as what it enables.
CapRE: Thanks Sharif. We look forward to seeing you in Denver.