 We're back here live in Las Vegas for theCUBE's exclusive coverage at siliconangle.com and wikibon.org's exclusive coverage of Amazon Web Services Reinvent. I'm John Furrier, the founder of SiliconANGLE. I'm joined by co-host Dave Vellante, co-founder of wikibon.org, doing deep research in the cloud, great work from Wikibon, storage, big data, now cloud. This next year, you're going to see a lot of cloud research come out and a great cloud coverage continuing on SiliconANGLE. Again, this is theCUBE, our flagship program. We go out to the advanced instructor listening from the noise. We're excited to have Aaron Peterson, SVP of Product Manager for IO.com. IO data centers, what's the official company name? IO. IO, the best domain name in the business IO.com. Just every time I see that domain name, I just fall out of my chair, I just love it. You think it's easy, but you do explain IO and people are like IO what? IO. IO. Who do you? Welcome, welcome to theCUBE. So you own the roadmap, you guys have a unique offering. You guys are in Arizona, get an amazing data center, huge facility there, awesome. But the world of the data center is changing. It's Amazon saying it's going to go away. Now, maybe for some businesses, but the on-premise data center is never going away for the median large enterprises, but it is changing. Software-defined data center is a term that's just now getting out there. It's being hyped up. I mean, it hasn't even reached hype yet. So I got to ask you, when you look at the roadmap, you're going from physical plant data center to the software model, what's on that roadmap? What's your vision for IO? So well, first of all, thanks for having me. Fired up to be here. Great day in Las Vegas, great show. So IO is a technology company that builds data centers and to be a little more specific, we build software-defined data centers. And to be even more specific, we build an integrated product across hardware and software. So from a data center standpoint, everything that's built from a data center standpoint is built with software optimization, software control, software in mind. The founders have a software background. We build everything with the idea of monitoring and controlling via software. So the reason why we have our modular data centers that we have is because it gave us a standard form factor that has allowed us to write software against that standard form factor and optimize it. So you guys have a big client, that's a public reference, Goldman Sachs. I also do in your IPO as well, I don't want to go there. Given that you're going to go public. But people are ordering data centers like servers. And that's a trend we're seeing that came out of the Facebook. And we hear in Amazon is doing their own thing. So there's a lot of trends going on that you guys are kind of vectoring into. One is a data center, give me a data center. I got a workload, I want a full data center around that. You also got to build your own. You have now DevOps, the high IQ DevOps guys are doing their own data center, they're building their own gear. So you got open compute, Amazon's a black swan in my opinion, they're off the charts in terms of, they don't really play with open compute. But generally people are going to start looking at open compute as an option. So these things are converging. Share with the folks your view of that market. And how should they be looking at that data center? So we view the market as, while we see cloud and we see cloud taking off, we also know that there will always be the opportunity for a hybrid solution. And there's going to be, there's many reasons for the need for a hybrid solution. But data sovereignty is a big, big deal for companies out there. Where is the data? Who owns the data? Who has access to the data? I just read an article yesterday on how to keep your data away from the NSA. And it was know your data sovereignty, know where it is, know who has it. So really every enterprise out there we see is requiring a hybrid solution. And we empower, through our data center technology, we empower the organizations to define a standard data center, implement that quickly, efficiently, software optimized, control it like they control other items in the stack. And allow them to do it in a standard fashion, in a fast fashion, in a cost-effective fashion. So Aaron, you're responsible for strategy. You've probably spent a lot of time thinking about the roadmap, talking to customers, trying to understand requirements and figuring out where to go in the future. So where are you going in the future? Where do you want to take this thing? Just in terms of, you talk about software-defined data center. John and I talk about this all the time. Everybody talks about software-defined but nobody's talking about the software-defined data center itself, the infrastructure. So what are you sort of seeing from customers? What are they asking for in that infrastructure? So really where we see it going is that data is going to be king. And right now we're seeing it with big data for applications and we're seeing it with big data with social media. We're approaching big data from a data center standpoint. We capture, we have over 30 billion rows of data regarding the operation of our data center using our standard products. And we have an applied intelligence group that is actually going through that data and determining the value, determining predictive modeling, determining where we can get value from that data and then offer that to our customers. But beyond that, the ability to see all the way through the stack from the data center infrastructure. So all the way, say from the utility to the generator to the UPS to the server, to the virtual machine in the server, being able to see all of that information in one feed is incredibly valuable. And let me give you a scenario where that value is. Today you can go out and procure enterprise software, say let's say for example SAP. And you can choose for SAP to either run on the back end with database with either DB2 or Oracle. With the integrated stack, you can run test cases and look at the power consumption when you run your test cases with DB2 and when you run your test cases with Oracle. For the first time you're able to actually see how efficient the code is from an operation standpoint. Because if you think about it, you're going to make that decision to buy that package up front one time, but you're going to run it for 10 years. How efficiently it runs needs to play into the equations. So one of the challenges of that type of data is that a lot of times the CIO doesn't see the power bill, right? So what are you seeing and how are you, I mean I presume that through a software defined interface you're able to actually make that connection. And it's happening more and more, but nonetheless, it's like infrastructure guys over here, tech guys over there, sometimes they talk. What are you seeing in terms of that trend and are you able to bring those two constituencies together? So common challenge we face every day. The way I draw it, if you put me in front of a whiteboard I'll draw a river and on one side of the river you have the IT guys and on the other side you have the facilities. And I need to find that guy who's standing on the bridge across the river. And it's not always the easiest thing to do. And the money's floating down the river out the window onto the sea. Yeah, exactly. So but the point is can a software defined data center as you described it and providing visibility on those metrics begin to create that bridge? It can, it can. And the question we like to ask people, they want to say I want to save money on my data center. So the first question we ask is do you know what you're spending on your data center? And typically they don't. Typically they can't go down and determine by application or by server what they're spending because they don't have the integrated stack, the data in an integrated fashion. So knowing what your expense is. Knowing tying, being able to tie it back, being able to charge back the appropriate expense is really where you need to start because if you don't measure something you can't improve it. We were talking to James Hamilton recently and he was a couple things ago he was showing some pie charts and obviously compute is a big culprit, right? In terms of energy consumption, right? Heat density is a real problem to the data center. But he was saying networking increasingly is becoming problematic. Where are the, in the blame pie of energy consumption where are you seeing the big challenges? It's compute. It's compute. The network while partially because of the way the industry is going, there's more and more network traffic. You need to be more connected. However, really where the power is being consumed is in the compute. Yeah, so spinning disk is relatively minor even though that's kind of going away too, right? Yeah. So toward flash. Yep. But so compute, so how do you think the industry is attacking and will attack that problem? Start by measuring, then what? So we're seeing some interesting things from the industry and interestingly enough it's being driven by the mobile market. So I know there's been discussions of moving to ARM and the efficiencies that you can get from ARM and I know there've been some claims that no matter how efficient ARM gets, you can't, we'll never overrule x86. But clearly with the density, so being able to densify your data center and then the technologies that we see coming down the path, the combination of those two, we're going to see significant consolidation in the data center for the amount of compute for the amount of space and power required. I got to ask you about a trend that's happening certainly Amazon's taking advantage of and that is the public cloud and they're winning big time and you hear everyone here, oh Amazon, who's going to compete against them? So in the public cloud infrastructure as a service, okay, I give them that. They won, are winning, continue to win. They're light years ahead of anyone else. IT is not the same. IT, it's just, it's cultural, it's apples and oranges, right? I mean, so the question is, does Amazon convert the enterprise? What does the enterprise convert to Amazon? So Dave and I talk a lot of time. Well, if you want to weigh in on that, be great. But also talk about what IT's like. What's the mindset of IT? Because they're not going to just flip their data centers, ripping replays, work lift upgrades, all the things that they talk about. What does IT need to do to be successful to go to that next generation data center? Well, first off, IT guys are control freaks. And second of all, IT guys have the not invented here syndrome. And disclaimer, I am an IT guy. Computer science background. There's legacy stuff. They have inherent baggage that they're dealing with. Previous investments, software, applications. I think it's going to need to be a generational thing. It's going to be, you know, there were managers and decision makers that did not let go of the mainframe. And it's going to be the same way with IT. However, I think as the generation gets used to never storing a file on a hard drive on their machine, never having information, you know, for themselves with storing everything outside. Once you see those people and they start proliferating in IT, I think we're going to see that, see IT go that direction. And I think you got, I think you're right on that. If you look at the Amazon success stories up there, Airbnb, Uber. I mean, these are young, green field, clean sheet of paper deployments. These guys are, you know, cutting their teeth in dev-offs. But you go to a large enterprise that's been around for, you know, a century. And in IT, in IT decades, they have a mainframe, they've got client service, they've got PC, lands, internet working, printers on the network, all that stuff. They're not compelled to just throw that away. So the question is, who will be the cloud there? I totally agree it's hybrid. So I 100% agree with you on that. But now, but they still get the scale problem. So you guys do a lot of work around the kind of providing infrastructure at scale with the containers. What does that business look like? Cause more and more people are dialing up more resources. What is that market like? I know you guys got a good backlog of business, but as you tweak that product, is it form factors, is it integration? Is it software, all of the above? Really where we're seeing a lot of interest and growth in that area is on the edge. And people looking to reduce latency where they know they need to reduce latency. They know they need compute in a location. They don't want to go through the entire exercise of finding a plot of land and constructing a data center when they can buy one ready to go off the shelf. Which is why our product line there is called IO Anywhere because you can put it anywhere. And you guys integrate a lot of power and coolant. Talk about that because we heard from James Hamilton, obviously power and cooling is something that they've done that's changed. What have you guys done to take advantage of some of those costs? I mean, obviously server is one big company. You got power, facilities, I mean, cooling, just subsystems, and then actual power itself. Those are key areas. How do you get those costs in line? So our modules, in our Phoenix facility, we operate standard data center area and then we actually operate in our modules as well. And we had a third party, actually our public utility, come in and do a study. And what they found was that our operating expense in the modular area was 40% lower than in the traditional area. And the reason being is in the modular area, we only cool what needs to be cooled. So if there's significant compute going on, which we talked about earlier, compute is the source of heat, then we cool that area. If there's not, we don't. You mean you're not chilling out the whole kitchen to refrigerate the sandwich? So you're talking the PUE metric on your pods is going to be much, much better than a standard data center, right? Can you share some, and I know it varies based on how people can figure it, but what are you seeing in the field? So in our multi-tenant data center area, we're seeing about one four. But once again, that's because we run to, it is multi-tenant. So there's different types of compute going on in there. So if you made it more homogenous, you could get down to one three, one two. Exactly, and it's all software controlled. That's the key. You can change the temperature by essentially logging into our dashboard and making a change in a field. And hopefully raise that temperature, right? Raise that temperature, lower your cost. And then we also have a product, our eco product, which takes advantage of free air cooling. So based upon the outside air conditions, it will actually use outside air for the cooling. If the humidity is where it needs to be, it can use evaporative cooling. And if the conditions aren't good for either of those, it will actually use mechanical cooling. And I understand Phoenix is actually a great place for ambient air business because it's not humid. Yeah, believe it or not, Phoenix is a good candidate for that. Denver, Colorado is a great candidate for that. So you don't have to go to Iceland to build a data center. No, you don't. We're here with Aaron Peterson, the SVP of product management for io.com, io, data centers. Or io. Or io. Just io. Just io. I want you to put a bumper sticker on the car that's leaving Vegas for re-invent. What is this show about? Share with the audience your bumper sticker tagline on that car, on the freeway. What does it say about re-invent this year? What's the bumper sticker say? Re-invent. Significant growth. Significant growth. Plenty of opportunity, but they're not going to get everything. Okay, that's three bumper stickers. Okay, we've got three in one. It'd be a square big goose on the window. Okay, we'll be right back with our next guest. Exclusive coverage with io here and Amazon Web Services re-invent. The cloud game is Rio. It's changing the economics. Hybrid cloud is right there on the doorstep. All this is all going on at the same time. It's a cloud collision here in Las Vegas. This is theCUBE. We'll be right back after this short break.