Compass Datacenters’ Sharif Fotouh Sits Down with datacenterHawk to Talk All Things Edge

In one of the best conversations about the edge, David Liggitt and Sharif Fotouh discuss the past, present, and future of the edge, from repurposed cell-tower “fiber huts” to never-before-seen data patterns. Watch the full video below or read the entire transcript. Enjoy!

David: I’m David Liggitt with datacenterHawk. And I’m here with Sharif Fotouh. He is managing director at EdgePoint at Compass Data Centers. And we were talking about the edge, it’s one of my favorite discussions ever. You gotta watch it.

I’m David Liggitt with datacenterHawk. And I’m here with Sharif Fotouh. Sharif, good morning.

Sharif: Good morning.

David: Thanks for joining us here on Hawk Talk number 33. So, really excited about that. And excited to get your feedback on the data center industry. Obviously, you have a very interesting focus around the edge and that type of, I would say, like, vertical in the way companies are using that or will use that in the future. So, excited to get your feedback.

Sharif: Oh, well, thanks for having me. And I wanna say before we get started, congratulations on your five-year anniversary.

David: Thank you. Thank you.

Sharif: It’s a heck of a milestone.

David: Thank you.

Sharif: And it’s been cool to see the story of datacenterHawk develop over the years.

David: Yeah, you bet. Well, I appreciate that. You know, let’s talk about, you have a really interesting background as far as some of the things that you were doing before EdgePoint and before being acquired by Compass. And so, talk about that, those maybe…your career path and how that really shaped…we’re talking about it a little earlier, but how that shaped the ability for you to go out and do what you’re doing today. I’d love to hear more about that.

Sharif: Great question. My career path is kind of if you imagine one of those Calvin and Hobbes comics with the, like, squiggly line, definitely not the most efficient route. But I think that variability kind of lended some interesting colors. So, I’ve been in the industry for about 15 years now.

I started working for a regional data center operator out of Austin, Texas. And I had a very interesting collection of roles there throughout the years, I was there six years. Part of it was operating, you know, retail colocation facilities. But part of it was also deploying for various web properties that was under the same family of companies in data centers all over the globe.

And so, throughout those five, six years, I kind of never specialized in any one aspect, but rather, had, you know, familiarity and experience with, you know, everything up the network and system stack, all the way down to power whips and space planning.

David: So, you got a really wide variety of kind of a knowledge base of the industry?

Sharif: Yup. Jack of all trades, master of none, right? In 2013, I started with Google on the Google Fiber project. And back then, it was, you know, very kinda early stage, a pretty small team. They had just had Kansas as their city. And it was publicly kind of known, it was just a trial or, you know, a test, right?

David: Yeah. So, they were putting their Google Fiber product in that city?

Sharif: Yup.

David: Got it. Yup.

Sharif: And so, I moved out to California to the Bay Area. And over the next five years at Google, I built and led the facilities and network deployment program for Google Fiber nationally.

David: Wow. What was that like?

Sharif: Actually, it was quite exciting. But at times, really challenging. Everything from, you know, mundane labeling standards to, you know, building templated footprints for various, you know, kind of peering functions, and that kind of thing, so you could kind of cookie-cutter it across the country. But it was really exciting, the mission was exciting, innovating in a space that’s, you know, traditionally pretty slow, telecom. And the public was, you know, really energized by the product. And so, you know, that kinda fed into the excitement.

So, that program at Google Fiber encompassed kind of a wide spectrum of facilities, right? So, everything from, you know, megawatt deployments within Google’s hyperscale facilities for kind of our back-end services, regional pops all over the country to effectively build your kinda national backbone network, you know, for acquiring other networks and peering. And then most pertinent to what I’m doing today, hundreds of these small prefabricated telecom shelters that were dubbed by the press fiber huts. And that name stuck.

And so what was really interesting about that fiber hut story is when I started, we had a very rough design that was effective, you know, 10 to 15-kilowatt footprint, you know, barely n+1 with a lot of single points of failure, it was a stretched cell tower shelter, right? They called the guys that are making shelters for cell towers sites, and they said, “Oh, yeah, we can build you a slightly bigger one and, you know, upgrade some components.” And then over the years, right, outages, failures, requirements for increased density, variability and footprint, we went from just kind of the typical PON gear, which is Passive Optical Network gear, went from just PON gear to including transport systems, including, you know, servers, and getting more advanced. We went through multiple iterations of that design on my team. And so, kind of what started as 10 or 15 kilowatts, you know, the last design we produced over there was over 50 kilowatts. And solidly n+1, if not 2n in some places.

David: Yeah, interesting. You know, you bring up the point about the iteration of the product. And, you know, it’s one of the things we were talking before just about when I started in the space back in, like, ’07, and just watching how, you know, the physical larger data center facility has been iterated over the last, you know, I don’t know, 5, 10 years. I was…just got back from Northern Virginia, and one of the things we were talking about is just the difference in design, you know, and how different obviously it is today than it was 5 or 10 years ago. And really, I think when that iteration begins and, you know, it’s really the product positioning itself for scalability down the road. And so you basically finished your time at Google. And then tell us about starting EdgePoint, and that process, and what was your mindset behind really wanting to get out there and go do that?

Sharif: Yeah. No, that’s a great question. And so, let’s see. Late 2016, the Google Fiber project, you know, effectively went on pause, right? There was a change in strategy. And the organization kind of halted the expansion plans. And I had this kinda moment of wondering, okay, like, you know, we went from racing down the highway to, you know, stopping, like, what do I do next, right? And what was really interesting, and this is kind of pre a lot of the discussions and conversations that are centered around the edge space now, you know, there were barely a couple articles and players in that space at the time. And it’s really, the initial premise was as somebody that was sourcing colocation all over the country, in tier-two markets, because Google Fiber was specifically targeting tier-two markets, right? I was surprised and shocked by how many markets were underserved from a colocation perspective.

David: Interesting. On, the secondary…in the secondary markets.

Sharif: In the secondary markets. And look, since that time, a lot of those markets have been solved by players in that space your EdgeConneXs and TierPoints. But what really surprised me was a study we were doing for a specific city. I think it was Minneapolis at the time. And we were going to have to pay an exorbitant amount to a colo provider there to add a second generator. They had one generator for the whole facility. We were going to buy a second generator for the whole facility. So, I could take my 150 kW footprint inside, right? We didn’t have, like, a huge footprint requirement, but we needed it to be…

David: Yeah. The redundancy, yeah.

Sharif: Redundancy, right?

David: Yeah.

Sharif: And I remember walking out of that conference call. My VP was with me. And I turned to him and I said, “You know, it’d just be cheaper to put a couple of our huts and, like, it would…”

David: Robust, yeah.

Sharif: “…it would be more robust.” And that was months before we went on pause. But that thought kept gestating in my mind. And so, when Google Fiber went on pause, I kind of locked into that and thought, well, hey, I’m uniquely positioned with this experience in producing hundreds of these shelters, deploying them all over the country, right, all the specific jurisdictional concerns and, you know, going to seismic regions or regions with snow load, you know, very humid and hot regions. And so, you know, I thought that ecosystem I’ve really tapped into and understood. And there is an option there to support areas where the capacity, the colocation capacity or facility capacity, is in an ideal location. And that’s kind of what turned into edge.

David: Sure. Interesting. So, for those that are watching that don’t have a good grasp on what edge is, you know, and you’re in it every day, and granular with the thought process, and the definitions, and things like that. But do your best to describe, you know, and I think you’ve kind of done that through your conversation with what you did at Google. But do your best to describe, you know, what the edge is and why you believe that there’s such a big opportunity, you know, in the future with this type of data center product?

Sharif: Got it. What the edge is?

David: Just a question we should all laugh about. Yes.

Sharif: Yeah. It’s, like, religion, politics, and the edge, let’s not talk about those three things, right? No. I mean, I like to keep things pretty simple.

David: Let’s go.

Sharif: So, on the simplest level, distributing infrastructure is not a new trend, right? When I started my career, you know, advanced organizations would take a footprint domestically so to say, you know, in Ashburn, Virginia, and, you know, maybe West Coast like Los Angeles, and maybe Chicago, right? And that was effectively the coverage you needed for those applications. Over time, we’ve seen, you know, a boom of secondary markets or other NFL cities that perhaps aren’t, you know, specifically acquiring subsea cables. But, you know, you look at your Atlantas and everything else, they’ve just boomed, right? That was the march towards the edge, right? And that same trend is continuing, right?

And so, what was secondary markets, we’re now looking at tier three markets, we’re looking at, you know, big metropolises like Dallas for example, if all your data centers are located in the center of Dallas or in the outskirts, it doesn’t do good for the majority of your actual eyeballs in residential earn, in customers, right? And so, that’s effectively, it’s just an extension of that same trend, right? As bandwidth demands go up, as application performance becomes more critical, you’re going to want to locate the capacity closer and closer to the user.

David: Yup. So that’s, like, a great point. Let’s talk about strategy and how companies…you’ve seen companies do that well or not do that well. You know, one of the biggest growth points in the data center market over the last, you know, three to five years has been the hyperscale market. It’s really taken the industry to new levels from a demand perspective, etc. But there’s also the enterprise user-base. You know, these are companies that traditionally have housed their infrastructure inside facilities of their own, they have come out most of them, half of them, into colocation facilities and now have some sort of hybrid approach with their IT infrastructure. So, you know, this edge idea and what you just described, there is an infrastructure strategy in play where some people are probably doing this well and some people are not. So, maybe talk about that, how have you seen companies do that well and approach their strategy well, maybe the enterprise user sector, how have you seen them do that well?

Sharif: Specifically in the enterprise sector, I think it’s important to, like, you know, kind of…there’s the Buddhist concept of the new mind, right, come into the problem with a clean slate. And that’s really hard for larger enterprises because, you know, you have, you know, years of backlog and technical data, you know, facilities of different ages and infrastructure that’s in different stages of its life cycle. And so, it really is hard to tackle that problem really as blank slate. But that really is, to me and from my perspective, the key to success for those enterprises, because there’s a lot of different toolsets that are now available. And if you’re thinking within the paradigms of 5, 10, 15 years ago, you’re probably not going to build an optimal infrastructure, right, or a topology.

And so, you know, specifically, you know, you see a lot of enterprises shift and kinda cloudify their footprint. And then very quickly, they realize there’s kind of a long tail of applications and services that don’t make sense in the cloud. They’re high bandwidth consuming, really expensive to locate far away. And so, you know, you see them kind of contract and reduce their, you know, enterprise data center footprint and as they cloudify. And then you walk through this building and there are 10 or 20 racks, right? And they still have to keep all of the infrastructure running for them, right? And so, that’s really the challenge that I think we have a new tool to offer. It’s like, hey, reclaim that real estate. You know, use it for whatever your primary revenue driving activities are. If you’re a hospital, use it for patient beds. If you’re manufacturing facility, use it for assembly lines. And we can put a facility in your parking lot that takes six or eight parking spots that are fully 2n, and redundant, and hardened, and free up all of that space, and all the operational costs of that large infrastructure.

David: Yeah. Let’s do a little deeper dive on the product. You know, when the Compass edge product, what is it and, you know, where can you deploy it? I mean, talk about the actual details around the product itself.

Sharif: Okay. I’ll try not to get down the rabbit hole with this one.

David: Come on, you’re good, you’re good.

Sharif: But this was one of my favorite topics.

David: Let’s go.

Sharif: You know, I’m really proud of what we’ve built at Compass. And the key premise is the edge is in about one or two facilities and, you know, something I’ve, you know, told our board multiple times, anybody can deploy one or two little shelters, right, a little data center. So, it’s not from an astronomical feat, right? The complexity becomes in mass and scale, right? And so, I have this kind of oh, crap moment when I was at Google when we ordered a new wave of 50 shelters. You know, we went through the normal pricing discussions and design discussions.

And finally, we were ready, everything was signed off. We issued the PO. And an hour later, I got an email, “Hey, we’re really excited about your business. And attached is a spreadsheet. Please put the delivery addresses for all 50 of these shelters and the dates you need them by.” And that was my oh, crap moment because I realized, well, now I have 50 little construction projects to manage, 50 little facility integration and commissioning projects to manage, 50 system deployment projects to manage, cabling, and racking, and stacking,

David: Hire 50 more people. We need…

Sharif: And they were all over the country, right? I mean, so it’s…And so, that’s the oh, crap moment. And everybody will hit that point, right? Anybody that’s in this edge space, you know, right now we’re kind of in a trial and nascent stage. But as the volumes increase and as the market demands grow, people are going to hit that point. And one of the breaking points there is those construction projects are rarely going to happen on the same time. And construction is ugly and, like, always unpredictable, right? That’s just the nature of beast. And so, you can try to fight that, right? But good luck. Instead, what we try to do is design around that. So, what means is, we need to be able to ship one of our facilities, one of the EdgePoint data centers to whichever location is ready first.

And here’s where the problem becomes, because what if one of your locations is in seismic zone four, and one of them is in a wind-rated area, right, where you need, you know, Miami-Dade County or something, right? So, all of a sudden, you’re playing this game with the factory where they’re going, “Okay, this shelter is ready.” And you go, “Oh, no, the site that needed the seismic shelter isn’t ready. We need the wind-rated one, or we need the one with the snow load rating.” and you literally are playing kinda musical chairs with the factory trying to consolidate your multifaceted construction schedule with their production schedule. And it’s, I mean, again, I lived through it and, like, messy, right?

And so, with our EdgePoint shelter, the key premise is that it’s a single consistent footprint that can go into any region. So it’s wind-rated, it’s seismically rated, it’s designed for, you know, high snow load areas, you know, hot and humid climates, wherever you need to place it in the country, it can go without any changes. And so, it allows a user to buy 200 of them and start 200 construction projects in just-in-time delivery as they come off the line, send them to whatever site is next.

David: Yeah. And physically, you can put them inside, outside, talk about that a little bit, I mean…

Sharif: So, technically, you put them wherever you want. With that said, we primarily designed it to be an outdoor shelter, specifically because we think from a real estate perspective, it’s designed for that highest bar, designed to be withstanding the elements, designed it to be a hardened shelter. And then, yeah, if somebody wants to put it into an existing shell, there’s no reason it won’t work. But on the other hand, again, that consistency in product is key.

David: Yeah. Yeah. Your comments about the production of these units is really interesting because we’ve seen the focus on supply chain at the big scale in the space over the last three to five years. I mean, all you’ll hear larger data center operators talk about hyperscale builds or they’re talking about supply chain, you know, how quickly can we deliver this, how efficiently can we deliver this, how, you know, inexpensive for the user can deliver it. And it’s interesting to see the supply chain conversation on the smaller scale, as far as the, you know, 100 kW, 200 kW range. Because I think what it shows is that speed of delivery and your ability to scale up quickly is super important in, you know, both sizes of requirements, and just speaks to the fact, I think, that the users today, the data center user expects or has matured to the point where, you know, they expect that solution to be delivered as quickly as possible. And it’s interesting to see you all work through the process of going, “Hey, now we know we can deliver X amount of these in this amount of time in these regions, and they’re all the same.” So…

Sharif: No, absolutely. And it’s about consistency and delivering, right? I mean, speed is obviously important, but at the end of the day, there’s somebody managing their application performance, they’re about to release a new feature, and they’re looking at their capacity curve and graph, and they’re going, “I need to turn on this new capacity the compute, the storage, its bandwidth, by Q3 or we’re sunk,” right, or, “we’re not going to release this new feature. We’re gonna lose against our competitors that are.” And so really, that’s the nature of the beast, and just understanding that you’re a tail on a very large dog for these organizations. And working around those constraints is critical.

David: Yeah. Have you seen any different when you think of, like, industry verticals, you know, retail, healthcare, technology, financial, those type of firms, have you seen any of those companies gravitate more towards this type of, you know, data center in this approach, or does each one of those have a different mindset around that? Have you seen the different industry types embrace what you all are doing?

Sharif: Yeah. No, that’s an excellent question. I think there’s kind of two different segments of verticals, or categories of verticals. So, one of them is kind of on-premise solutions, right? So, I’m a university, I’m a hospital, I’m a manufacturing facility, I’m deploying more connected devices, there’s more data being generated, that data doesn’t all need to be backhauled, right, you know, video cameras, your surveillance cameras around your facility, that doesn’t need to be cloudified, right? I just need to store 30 days of retention and then throw it away. And so, the on-premise solutions is in one category, I think the other…And honestly, you know, as I look at the edge space in general, I think it’s often ignored. A lot of people are instead of gravitating towards the second category, which is that kind of wide mesh, you know, I’m going to deploy 5, 10, 50 of these around a city, and, you know, there’s usually graphics of, like, autonomous cars or somebody with a VR headset involved, right?

And, you know, definitely at the base of a cell tower, it has to be at the base of a cell tower. And that’s in its own category. And that opens up a whole slew of verticals, whether they’re hyperscale companies, MNOs, Fixed Network Operators, anybody could technically play at that kind of edge mesh network space. But to be honest, like, while everybody’s attention is there, I think there’s a lot of applications that we’re assuming will require X, you know, milliseconds of latency or microseconds of latency. Those applications haven’t even been developed, much less adopted, much less those requirements fleshed out, right? And so, I’m not saying that trend isn’t going to exist where performance needs to improve, but, you know, from my perspective, we’re kind of assuming a lot as, you know, facility people and infrastructure people, as far as what the future will hold.

David: Yeah. You mentioned 5G. And, you know, talk about the impact that the growth of 5G and the maturity of 5G will have on the edge market. And what opportunities does that create for your team at Compass?

Sharif: Oh, great question again. No. So, everybody is excited about 5G, right? It’s one more G better than what we have.

David: [Inaudible 00:21:09].

Sharif: No. I mean, it’s obviously exciting, especially as mobile traffic keeps growing, right? And so, you know, we’re now hitting the constraints of what our mobile devices can support. And we’re seeing breaking points in stuff like jitter and latency on the performance on 4G networks. So, 5G is super exciting. With that said, 5G in and of itself doesn’t demand edge computing, right? So, 5G is vehicle, right? And so if I say, “Hey, I’m going to give you a car that’s twice as fast as your current car,” does that mean you’ll get everywhere in half the time? Well, no, right? You’re still gonna obey the speed limit, you’re gonna be at stoplights, right? Now, if I say, “Hey, if I give you a car that’s twice as fast, but there’s a family emergency across town,” well, yeah, then you’ll probably be. So, that’s the need there, right? That family emergency is the need or the application. And so, sure, 5G will open the door to a bunch of needs and applications, but which one that will be, it’s still to be seen.

David: So, that’s a great tie-in to what do you feel like the needs are that are coming, that will help drive that? You know, so, say 5G is in place, it’s efficient, you know, what are the needs that you all look at and go, hey, once these hit the way we anticipate, it really changes things for, you know, the market?

Sharif: Edge computing…So, first off, we really like to take an application-centric perspective, right? So, let’s focus on the top of the stack, and then trickle those assumptions all the way down, right, rather than starting at the bottom. So, when you think about the applications, and you look at the trend, I mean, and we’re gonna skip past, or I’ll skip past all the, like, you know, surprising statistics over, you know, in the year 2022, data usage is going to be a kazillion petabytes, and, you know, like, sure. But we can all agree, we’re not gonna be using less data, and it’s probably a safe assumption we’re going to be using more sure. What’s really interesting about our data usage patterns, however, isn’t just the quantity. What’s specifically interesting about it is that the type of data that we’re using is changing, it’s evolving.

So, if you think of the internet 10 years ago, or even 5 years ago, those data patterns were largely a broadcast network. The majority of data was video data that was getting sent from central sources and down to our TVs and devices, right? And before that, it was music, and before that, it was, you know, email or articles, right, a broadcast network, duplicating, you know, newspaper, radio, TV, right? Well, now, we’re seeing interesting data patterns that aren’t going top-down, they’re going sideways, they’re going up. If you think, you know, my little surveillance cameras, my Nest cameras or whatever, is generating as much data as a TV station, you know? Like, I’m effectively a broadcast station at this point. I’ve got five channels, you know?

So, what an interesting traffic pattern. And now, the question is, well, does that traffic necessarily need to go over to Ashburn, Virginia or Phoenix? Now, chances are, I’m down the road in my office, and I’m just checking that, you know, UPS delivered my latest package, right? Why is that data getting backhauled across the country, right? When you order an Uber, why is that data getting sent out of state? You can’t even order an Uber from out of state, you’re definitely ordering one nearby, right? So, those are local traffic patterns that we’re seeing with the data and the data going in different directions. And that, to me, is what’s really going to drive a shift in the facilities and network architectures that support those patterns.

David: Yeah, that’s interesting. You know, one of the interesting things I read about, this is probably three or four years ago, but it was talking specifically about how YouTube and how when that platform arose, how it really, you know, changed, it gave everyone the opportunity to, now you’ve got so many people producing data, and it’s a similar…you know, you mentioned the Nest cameras, but it’s like a similar approach, now you have this technology pieces that are producing the data. And if we put them into use from either a business perspective or a consumer perspective, the amount of that data is just exponential. And that’s, today, let alone, you know…

Sharif: Whatever, 20-whatever, right?

David: Three years from now, five years.

Sharif: And it’s not just that, that old broadcast model was largely supplemented by local caching, which was great, you know? Netflix has very efficient caching programs. And their Open Connect program is great at intercepting a lot of those requests and serving video very locally, right?

David: Yup.

Sharif: That’s how their business model has managed to survive. But, again, when you’re talking about broadcasting your videos, you know, to your grandma across the country, or your extended family, all of a sudden that local caching is completely useless, right? And so the way we’ve designed our systems, infrastructure, and topology, it needs to evolve.

David: Yeah. How has the hyperscale growth and what’s happened over the last several years, how has that impacted the edge space? I mean, has that had a positive effect on the opportunities, a negative effect, neutral, maybe too early to tell, what’s your thought on that?

Sharif: It’s a complex ecosystem. Like, I don’t know that I’d go so far as to say this is directly causing, you know, that. But there’s a lot of correlations you can tie. So, on one side, you know, and I’m going to get a little more specific not just hyperscale but specifically the drive towards cloud, has pushed consolidation. It’s pulled out capacity from local areas. And has opened an opportunity for a solution that will effectively let your capacity get closer to users, right? It has driven a little bit of a need. On the other hand, because those hyperscale cloud players have such an immense network and also, you know, their scale, it also gives them the opportunity to potentially get into the edge game themselves and just offer their platform closer. And you can see some of the early kind of band-aids of the edge need. You know, AWS has their Greengrass project if you’re familiar.

David: Okay. No, I’m not familiar with that.

Sharif: So, this is like, and I hate to call it band-aid, but that’s effectively what it is. It’s, you know, an IoT device, an IoT aggregation device that they’ll ship to your locations effectively off-net, right? Your manufacturing facility out in West Texas, you have all these IoT and automated devices, and you want to run everything within the AWS environment, they’ll ship you a little box, right? And today, it’s pretty rudimentary and corner case but it points to the requirement and the need. And so, as that need grows, as those factories get more automated as more data grows, well, that little box is gonna grow and before you know it you’re gonna have a few racks of it.

David: Interesting. For the people that aren’t as familiar with the edge discussion, you know, there’s a lot of…some people are familiar with cloud computing, we’ve talked about edge, but there’s also terms on fog computing. So, help demystify the edge versus the fog from your perspective.

Sharif: Yeah. The key to the edge markets…

David: And sometimes I feel, like, you know, not just the data centers, but the technology industry, just thrives on confusing people, you know? It’s like, “Hey, how can we…” Anyways…

Sharif: Oh, no, no. You know, I feel like as the technology industry has grown, marketing has become just as key as the actual technology. And so, you know, Internet of things turned into, you know, internet of everything, turned into, you know? It’s just…

David: Is there a difference there?

Sharif: Yeah, no. So, fog computing is effectively an implementation of the edge that’s effectively a shared resource pool with multiple distributed facilities. So, it’s…the simple way to say is all fog computing is effectively edge, but all edge computing isn’t necessarily fog computing. I could put a single edge footprint that’s not connected to a bunch of them, that those workloads are solely homed at one location, that’s not fog computing. But there are applications where you could see benefit to shifting workloads to nearby, you know, facilities if you have a mesh network of facilities. And what’s primarily kind of enabled that is in the network space, A, like, networking high bandwidth, like, high capacities gotten enormously cheap, right? Like, the network devices themselves and the capacities they can handle, I mean, I remember when, like, 10 gig optics were out and it was, like, insane. And, you know, now you’re talking 100 or 400 gig. And now, people are pushing terabits.

And so, that with Network Function Virtualization, or NFV, has effectively allowed the network to be the backplane now. It’s the system backplane. And so now, you can connect storage units, and, you know, GPUs, and CPU units, and build a virtual system, and they don’t have to be physically located in the same spot. And now, that opens up kind of the concept of fog computing is I can kinda sprinkle my capacity around and poach from it as I need.

David: So when a user has a need and they, you know, either they approach you all or other edge companies, I mean, let’s say, they’re more enterprise user, and they might have a need for one of these. But let’s say they have the need for 10 or 20 to solve a problem, how quickly can those be deployed based off of, you know, supply chain, and where they send them? I mean, what should a company be thinking about how quickly you could set up a complex system of, you know, these different deployments?

Sharif: No. Complex system, I’ll absolutely echo that, right? It’s not just about manufacturing the facilities in time, you have to think about the program holistically. And that’s really, I think, where our experience at Compass lends an advantage because it’s not simply about spitting out, you know, little data centers out of a factory, you have to work on your site, you know, your whole real estate program and site development program. You have to ask all the questions that data center people generally leave out, which is, how are you going to accomplish, you know, system integration once the facility arrives? How are you going to commission these remote facilities? How do you ship $5 million of IT to a location where you don’t have a facility, a non-addressable location?

And, like, look, there’s experiences of, you know, guys on my team towing an 800-pound router up a dirt path up a hill to get one of them, you know? And this is, like, a million-dollar piece of gear, right, with a little, like, hand truck, a U-Haul hand truck, right, like, the rented dolly. And so, those are the areas that anybody that’s going to be successful in the edge space really need to focus on, not just focusing on their simple, you know, facility design, but zooming back and looking at a holistic solution that solves for those customers of scale. Because calling them and going, or setting them that spreadsheet and going, “Let us know where you want them delivered.” Well, now, you’re commoditized product, you’re not really an ecosystem product.

David: Yeah, sure. And it is so strategic, you know, with the end game of serving the user better, to your point. I mean, and those are really good questions. We should make sure that we log those in the show notes just on how companies are thinking through their strategy with what they’re actually deploying. And, you know, I think what this shows, and in your experience certainly proves, is just as the data center user is maturing, there’s just…it’s a really exciting time in our space because it’s giving, you know, your work and others’ work in this space and…is giving people the option, companies the option to build more efficient systems and ones that serve their business better, that serve their end-users better. And so, I just think it’s a really exciting time. And I’m fascinated to think about the next three to five years and what this space will become.

Sharif: No, absolutely.

David: And maybe talk about that, just the future of EdgePoint and the Compass EdgePoint product, and, you know, why you’re excited about the next five years in the space, and what you think the opportunities are.

Sharif: Yeah, I think, so, first, you know, as kind of a self-professed geek, right? The next three to five years are going to see just a really interesting array of applications that I’m very selfishly excited in because I hate sitting in traffic and driving my own car, and, you know? So, like, A, just I’m really excited about all the gadgets coming out, right? Like, life has just gotten better with technology, right? Like, I can monitor my house, I can unlock my door, like, I can do so many operations that, you know, used to be tedious and, you know, very manual before, right? So, A, the applications are exciting. And that just excites me as a technologist. Second, it’s really being able to leverage our experience and support those users in understanding kind of the journey that needs to be traveled as their application grows and needs distributed infrastructure.

Again, the lessons I learned over the past five, six years being in a program like this, I can’t wait to leverage those for our customers, again, not because we’re necessarily the smartest people here, but the scars on my back, I’d love to turn those into lessons learned and get a positive outcome, right? Lemons and lemonade. So, I’m really excited to be able to, you know, work with our customers and new potential customers to really lean on that experience, build a program for them that is holistic, and show them the advantages that we can bring to the table.

David: Yeah. Well, Sharif, thank you for sharing your insight with us. And, you know, for those that are watching, this is a fascinating discussion because it’s very much today, but it’s very much looking out at the future of where things are going. And so, I just always appreciate people that are willing to share what they’ve learned through their career path, through where they are now. So, I’m certainly excited about what you all are doing and it will be fun to watch over the next several years.

Sharif: Oh, thank you.

David: You bet.

Sharif: It was a pleasure.

David: You bet.


About Sharif:
Sharif Fotouh is the Managing Director of Compass EdgePoint and an ex-Googler. Fotouh is responsible for the Compass EdgePoint’s edge data center solutions as part of our comprehensive core-to-the-edge offering to customers. He is recognized across both the information technology and the data center industries as one of the preeminent experts on edge computing. He has more than 10 years of tenure leading large data center and technology teams, including founding and leading Google Fiber’s national network facilities and deployment engineering program.

If you’re interested in seeing how EdgePoint and Compass Datacenters can help your organization make the most of your data center needs – now and in the future, reach out to us and someone will get back to you shortly.

We Build for What’s Next

Learn More › No thank you.