 Las Vegas, this is SiliconANGLE, the theCUBE, our flagship program, we go out to the events, expect a signal from the noise. I'm John Furrier, the founder of SiliconANGLE. My co is Dave Vellante, and our next guest has been here to talk about the future of the data center. Brian Fanzo with IO data centers. IO data centers, welcome to theCUBE. Thank you. So Amazon says we're going to replace the data center, kind of yesterday they were saying. The data center is going to be gone in the future. Okay, small business, maybe law firm, small group, maybe use the cloud, but data center is not going anywhere in the future. As Dave and I were talking about yesterday. So let's talk about the software defined data center. So talk about IO, your company, and what you guys are doing around here, cloud and the data center. So from IO's perspective, you know the data center is the foundation of any cloud, the cloud has to live somewhere. So at IO we have a global footprint of data centers that we've owned and operated. We went from a co-location kind of product service type company to now we have modules that are manufactured in US and Chandler, Arizona. And we actually moved from the modules to adding the software, which we call iOS, which is a data center operating system software. And we call software defined, I think everything seems software defined these days. I have a software defined watch, but software defined data center, but really it's a software controlled data center. And we actually are able to provide that control because of the modular footprint that we actually have. So we have modules in Singapore, as well as Ohio, New Jersey, Phoenix, and the ability to kind of scale as well as coverage with software in the data center where we're taking the next step, kind of where everyone else really wants to go. And you can't have a cloud without a data center. That's for sure. We met with James Hamilton yesterday, VP Distinguished Engineer at Amazon. All he's talking about is scale, scaling data centers. Obviously they build their own data centers. We talked about open compute, which is the trend towards essentially people cobbling together their own hardware. But at the end of the day, it's a software paradigm. Amazon is showing the way that people want pre-defined software application workloads infrastructure. And everyone wants Amazon on their data center. So talk about that trend. Talk about the build your own customizations. Do you have some big name clients like Goldman Sachs? I mean, these are not small firms. I mean, these are large data centers. So talk about the trend towards this movement where we're going beyond the rack and stack gear too. I need to have pre-configured, pre-tested hardened infrastructure that looks like cloud, feels like cloud and is cloud but on premise. And I think that's a good point. I was participating, I think they are calling it hybrid IT or hybrid cloud. But I think to be actually a hybrid on premise, you have to have a data center on premise. And with IO, we have modules that are actually installed at our customer sites. So some of those customers have them in their own data center, in their own site. We have modules that actually can be dropped in any location. And part of that for the on-prem, you have to have software that ties the data center together. You can't have a hybrid solution with no real software that bridges a co-location or a decast data center that you have an on-prem data center. So to make the actual hybrid cloud or hybrid IT work, you have to have software that bridges that gap. And that's kind of where the software to find data center comes. If you are able to actually monitor and manage power and IT infrastructure all the way up the stack across not only what you control as their private cloud or on-site, but also what you're paying for at a decast location or even your data center 1.0, traditional environment. That our software iOS kind of bridges that gap and kind of enables everyone here. Amazon is pushing the envelope, but if you are going hybrid, you have to have software that kind of ties the two together for sure. So Brian, John and I have seen your facility, very impressive physical security. I mean, the first thing you're greeted with is basically a roadblock. You can't go better than that. Hard to go in and out for lunch sometimes. Which is kind of what you'd like to see in a modern data center. But talk about what's different about when you walk into one of your data centers, what's different between that and what you might see at the run of the mill data center and whatever, New York City or any major city or out in the rural area. So from our data center footprint, we have multiple locations like I'd said and the Phoenix site is our headquarters where we're kind of based out of and the idea of security is a big deal. I think a lot of the conversation we're hearing here is security is definitely something that we have to be aware of and what we're able to do with our software for the data center is we're actually able to control the infrastructure but we're also able to provide security elements. So when you walk into our data center, you're greeted by man traps in your traditional raised floor data center security but when you go into the modular section of our data center, which is we've built out as we grew through traditional, we are either on dog food, go to say. So we moved from traditional raised floor to actually modules that are located in our Phoenix facility. The security layer has now taken another step and I did nine years of cyber security with Department of Defense and one of the big things with security is you want layers and the module gives us that layer as well as the ability to actually control those that layer with the software. So the software to find our software, we call it software controlled or intelligent controlled platform and when you go to our data center, it looks different because now even though you have access to the data center you've gone through these three security now you can actually block off your security per module and the other nice part is it's separated between your IT gear and the infrastructure support below. Well, I mean, I was impressed. I mean, I've never been at the data center like that. I've never been at Amazon's data center, obviously. Right. I don't know if anybody's ever been at Amazon's data center. We don't know where they're at for some of the locations as well. The first thing you notice is the size. I mean, it's enormous. So what kind of square footage do you guys have globally? So globally we're over 2 million square feet. So we're just right around that number for... So that puts you at the same scale as an Amazon or a Google or a Microsoft. Is that fair? Yeah, we're playing that square footage wise. And we kind of joke because the square footage is a lot of the way that people that talk about data center 1.0, everything's square footage. But we actually take it because of the density we're able to drive with our modules. We actually have higher density IT than that 2 million square footage. If you took it, transferred to another company that just had traditional race floor. Well, that was the other thing. So it was very large, but also like spotless. Now some data centers I've been on are very clean. Some aren't, but yours obviously are. Very well organized. And the pods were very, very dense. And we also saw the manufacturing facility for the pods as well. And John and I, when we were in Barcelona a few years ago we saw the HP pods. These are quite a bit different. They were actually roomier, but you would argue more dense, right? And we talked to some of the guys who were building them. They came from HP, but so talk about that a little bit. So they're manufactured. They're not your shipping containers. We don't, they're designed and manufactured. They're in Chandler, Arizona. And I like to think of it as everything's fine tuned. We have a applied intelligence group that's out of San Francisco. And they were teamed up with McLaren, the race car company, which I think, kind of pushing the boundary of kind of analytics, real time simulation. We are working with, I was talking a couple of weeks ago with the McLaren group. And one of the things they do in a race car for optimizing a, every single square inch, every square second matters for a race car. We're taking that same analytics and putting it towards the data center. But not only software, but we're doing that for the physical, what you see, we call our IO anywhere modules. And so with those modules, we've engineered every square inch of the module. So when you go inside the module, it's your, 18 cabinets is our traditional data module. And it has a four foot cold aisle, three and a half foot hot aisle. But beyond that, we have sensors. It's close to about 800 individual sensors in that module that are controlled and managed through our software. So, the idea that every single square inch has been manufactured for a reason, it's different than when you go into your traditional data center 1.0 environment and it's your air conditioning, roof to ceiling. You could take a Kmart and a Walmart and think about how do you manage and control the infrastructure that's supporting that? You don't really have very much granular control in our modules because we've engineered them, gives our customers as well as us for our co-location, our Dcast footprint, gives us that ability to control. So what do you consider the useful life of your data centers? Well, so right now we have kind of a big number for us is we actually hit over a million operating hours for our software with our modules. And the reason that's a big deal is because we actually now have that much data to actually go back and actually analyze, are we over provisioning, under provisioning? Can we drive higher density by changing the way that the air comes from our air handling units below the floor? And everything, customization seems to be a big trend here. We were talking about that a little bit about the ability to actually take and do customized individual solutions. And at I.O., you don't have to have, you can build out, scale it out because you can have four air handling units underneath your module if your density's not very high to begin with. But as you're increasing your density instead of having to build a new data center or even buy another module, you can actually increase the amount of air handling units that you have in your module by just plug and play components, and now you can have higher density in that same module without having to kind of procure something else. So from a financial, from an accounting standpoint, you're depreciating your data centers over a long period of time, right? I mean, essentially you're in the real estate business in part, right? Right. But I've seen a lot of data centers today and you talk to the CFO and the company asks, how long are you depreciating this asset? And it's like, well, 30 years. But a lot of these data centers aren't going to be around in 30 years, at least in that current form. Correct. So what's going to happen to all those data centers that are out there? I think it's funny, because the data center, I think it's real estate, it's been real estate. But we were at VMworld and a couple of people had come up and asked, what's IO doing at VMworld? And we aren't a data center company where you're just a real estate company. We've taken the initiative and the drive to kind of go modular. So now you can have a module on site, you can set a module outside in Arizona and Alaska. We joked that one day they can take modules and run up the mountain if they needed to. And really what we're changing the data center from, it's just something you buy, you depreciate, and then you have to either scrap it or build from scratch. And we're actually making it something that every single time that you change an IT. I joked that the iPhone six years ago, we didn't know that you needed, because the iPhone didn't exist, we didn't know that you needed that much mobile footprint, that much compute to support that. And being able to scale kind of eliminates the walls of the data center because of our modules. And that's, I think the future of the data center is really, it's no longer square footage won't matter. And really the location won't matter because that's with our software. And we kind of talked about hybrid IT, hybrid cloud. The ability to actually, using our software, the bridge your location. So we have a data center that's in Phoenix, Arizona, and we have one that's in New Jersey as well. You're actually able to not only manage and control both of those locations from anywhere, but you can actually shift data load between the two of them as well. So the location of your data center will soon not matter as well. And that kind of gets you out of the real estate conversation as well. Ryan, talk about the notion that Amazon's putting out there is the internet operating system. Because essentially what they're doing is talking about notifications, talking about all the stuff that we were joking yesterday in a realistic way that's just like a mainframe. Basically globally distributed large scale computer called the cloud. That is the operating system model. So you guys have a unique approach. Talk about how you guys look at that. Fundamental holistic picture as an operating environment. You got software, data centers, not just gear. Talk about what it takes and how you guys are rolling out because you guys have an innovative solution. You got huge clients buying data centers basically from you guys. You ship them to them. But it's not as easy as just rack and stack and gear. There's a big software element. That's kind of what Amazon's talking about. IBM was talking about it last week at IOD. A lot of notifications, a lot of scheduling. This is the elastic infrastructure model. So what's your take on that? How do you guys fit into that picture? And I think that's kind of a perfect tie for us. And the cloud environment is interesting because now what companies, especially enterprise companies are deciding is they're no longer, they're shifting the responsibility. They originally said we want to control everything. Everything has to be on our premise. Then they move to the co-location or your D-Cast type environment. And when they move to that, they said okay, I want to give you some control but stay out of the rack. And now we're at a level with the cloud and I think everyone here is, we don't want to have any control or anything that you have tied down to the infrastructure. And we know the data center won't go away because it's the foundation of anyone's cloud. If you don't know where your actual data center is that's supporting your cloud, how are you confident that you're uptime? Because if the lights go out in your data center, it doesn't matter what cloud provider's sitting above it, you're still going to be lacking service for your customers because when it comes down, we all want connectivity 100% of the time, 24 by seven, and the data center matters for that reason. And with our software, the operating system actually gives our customers the ability to actually shift that responsibility to something they log into. We have a mobile app that allows you to view updates and notifications, but viewing, and I think that's even kind of something that we're working with McLaren to, okay, if I get a notification that says that this, let's say, the cold aisle temperature is above the threshold that I asked, you want to be able to actually control to change that. Giving me data that doesn't allow me, giving me the answer, that's kind of like asking the question without having to answer. Well, at IO with our software, we've designed it to where you can actually control that infrastructure and now make the changes in the software without actually having to go to the physical data center. Okay, so walk me through the use case, okay, of a customer, okay, because you guys aren't dealing with mom and pop, talking about like Goldman Sachs, huge scale coming. So walk me through a typical use case of your customer base. Obviously, you know, we've talked with Facebook, we've talked to Google, but Facebook basically has a procurement where they're just buying so much gear. I mean, the app guys say, I need this and they just basically ship boxes, they build their own. But that's the enterprise that's going to move into a different way. So take me through what a customer looks like, why are they calling you and what you guys do. So with our customers, what we provide is we provide kind of the ability for them to have their data center any way they like it. It's almost like, you know, Burger King, you get it your way because for us, if you have a co-location or you have your own data centers already or maybe you have a couple of data centers in different locations and then you want to have one on-prem as well and then you want to have one in our facility, say in New Jersey. You can actually with our software control all of those different locations, unique environments. You can actually control and monitor all that software from one global single pane of glass. And really our customers look at that as, okay, I can now scale without having to get rid of everything I've already built and kind of move from the traditional. We talk about the cloud, everyone's trying to move and migrate up the IT stack. Well, you can migrate your data center the same way you would migrate anything else that you're turning into virtualization and so on. So the other thing that those guys talked about in the analyst briefing was the cost of computing. Servers number one, storage, networking, the gear, mainly servers. And the second one was power cooling facilities and then third power itself. So those are like the top three. And then the rest were like, actually networking was like lower than that but networking gear and then it was other. So servers, cooling subsystems and then power. And that's our bread and butter. So I'll talk about how you, they're saying, well, hey, we take that off the table for you. I mean, that doesn't really, they don't really do that per se, but still they bundle it in. But you guys also have a unique approach because those are huge. How do you solve those problems? So we, so for our customers, they actually have the ability to drill into the operating system and see the power usage all the way down to the individual server. But you can also scale it back to the cabinet that you're running or to the module or to the entire physical data center. So the power cooling, I kind of joked sometimes that if you don't really understand the power, if you can't look at a matrix in real time and tell me the power that your data center, you're paying for, how can you, how do you know that you're getting what you're paid for? And a lot of the power that's out there, you pay for it as kind of like you're, for us, we want to give you the option like a power company. You can log into the software and you can see real time what the power, what the cooling is running in the data center supporting your infrastructure. So is that what the software defined data center is is to you, the ability to visibility on the data center infrastructure and be able to control it? Why don't you talk about that a little bit? It's the intelligent control. It's beyond even really just the software. It's the ability to actually take those million hours. We have 30 billion lines of data in our data center, in our SQL database. We can actually take that data, applied intelligence group out of San Francisco, then run analytics against it and then using our software actually control your environment. I focus on the word control because a lot of times in your car, if the check engine light came on after the car had already overheated, it's not very valuable. But ideally it comes on before you take it, so you can take it and get it fixed. And with our software, we want to intelligently control it. So we're going to fix the problem before it actually causes you downtime, which we all, that's the foundation that we're all worried about. We want to eliminate the downtime and concern of our customers. Yeah, okay, software defined data center. Very few people talking about the data center, John. I mean, shipping data centers is a good business. I mean, obviously the data center's not going to go away. Certainly with internet of things, the notion of a data center, although the gear might look differently, builds your own, but certainly, there'll be more and more IP devices inside a company. So certainly we're bullish that the data center's not going to go away. It's just going to change. So you guys are doing a good job there. io.com, a commentary on CrowdChat is the best URL on the business. io.com must really penalize you guys for Google juice, for page rank. I mean, it's like a two letter URL. Easy to remember. Yeah, pretty nice for email addresses, that's for sure. I wonder if Google is knocking your door for Google IO conference. IO data centers, congratulations. Brian, great to have you on theCUBE. We'll be right back with our next guest after this short break, live exclusive coverage from Amazon's re-invent conference, all about the cloud, all about software, all about large scale. We'll be right back.