 The Cube at OpenStack Summit at Lata 2014 is brought to you by Brocade. Say goodbye to the status quo and hello to Brocade. And Red Hat. Here are your hosts, John Furrier and Stu Miniman. Okay, welcome back everyone. We're live in Atlanta for the OpenStack Summit. This is The Cube. This is our flagship program. We go out to the events, we extract the signal from the noise. I'm John Furrier, the founder of SiliconANG. I'm John Stu Miniman, the analyst at wikibond.org on cloud. Our next guest is Aaron Peterson, executive leader, global services for IO, IO data systems. You guys, huge success. We at OpenCompute, welcome to The Cube. Thank you. We at OpenCompute, you guys, the CEOs on stage, provisioning up like stuff, look like a demo, but it's actually live stuff. So I got to ask you, this is a DevOps show. And you guys are essentially a DevOps company in the sense that you have on-prem, huge data centers, and now a cloud, enterprise cloud strategy. So first I got to ask you, what is your definition of enterprise cloud? So our definition of enterprise cloud is cloud that can be consumed in the way, any way you want it to be consumed. Whether you want to consume it as a public cloud, you want to consume it as a hybrid solution, you want to consume it as a private cloud, and wherever you may want to consume it, whether you want to consume it at someone's location, your location, anywhere in the world. You know, one of the things that I've always said was that VMworld George Leskman was on theCUBE and he was great at theCUBE. He won't come on theCUBE now during the quiet period. For sure, I know you guys are going public. But the thing that Pat Gelsinger said, the CEO of VMware was, you know, Software Defined Data Center still has a data center, right? And so one of the things that I think that's very interesting about what you guys are doing is you have a very specific solution for large companies who are doing the hyperscale in the enterprise for on-premise expanding on the facility side of it with the containers, the real containers, not like Docker containers, but like, you know, getting infrastructure up and running and now a cloud. So explain that concept, okay? Explain that solution and why is it so popular? Why is it getting so much traction? So the way we like to explain it is it's a platform and our platform has grown and hence why we're at OpenStack. But essentially our platform consists of our hardware, which our hardware is real data center hardware, not just hardware that runs software, but data center hardware that provides power space, cooling, mechanical, electrical plumbing, that hardware, the infrastructure that makes everything possible. It consists of our hardware, it consists of our software, which completely controls that and we have since taken it up into the IT stack where we have recently rolled out and we rolled it out in February, our cloud reference architecture, which consists of open compute hardware with OpenStack software running on top of it. So what do you say to the folks out there? We had one guest on the queue that said, oh, this opens, but I always ask the question, what's the future OpenStack? And then they always answer, oh, this is going to completely change the game on server economics, meaning it's going to decimate that server business. You know, the HPs, the IBMs, IBMs already sold their low end server business. So how do you address that? I mean, do you look at that? I mean, because what you guys are doing is essentially catering to a large footprint issue. You see Amazon making their own gear. You see Google making their own gear. There's a trend towards turnkey data centers, right? So the question is, how does OpenStack impact that argument that it's going to decimate the server business? So the best way I could explain this is the way we got to where we are is our largest enterprise customers used to come to us and say, hey, I need to procure and quickly roll out data center capacity. And they need to do it in 400kw, 800kw increments. We met that need. We created those products. We've been successful in that business, but it changed. They've now come to us and said, I need to spin up 3,000 virtual machines and I need to store 18 petabytes of data. They don't tell us. You're like data center on demand. Yeah, they don't tell us. I need an HPDL380 with this. No, I need to spin up 3,000 virtual machines. I need it on an open platform and I need it ready to go. So you're riding that trend. So you're basically saying, look, there's obviously going to be servers. Servers never kind of go away with what is a server. So you're saying it's not about the server per se. It's about the data center. Yes. Yes, and we see compute going to much more of a fungible type solution. If I can spin up a virtual machine, if I can do my workload, if I have the ability to do that and get all of the support behind it. So that's the key here is I don't have to worry about a construction project to allow myself to have the capability from a power space and cooling standpoint with IO and the Intelligent Control Platform, you can drop that down. You have everything that you need ready to go and it's completely software optimized. It is a true software defined data center, including the data center. Yeah, so Aaron, we're hearing everybody talk about how open sacs going to change the economic models of what's happening. We know that IT just wastes so much of their budget on just trying to optimize things and keep the lights on. But when we hear companies talk about, even when they're adopting new technology, building a data center is something that nobody wants to do. David Floyer, our CTO, wrote an article talking about mega data centers and said, even if you look at Amazon, they don't want to build another data center. Why is that such a pain point and how do you guys help address that? You know, it's interesting. People don't want to build mega data centers but they want to build mega capacity. So if you're lucky enough and you go viral and you have to have the physical need, you will build a large data center. I think with IO, the big benefit you get there is you can do it in increments. So you can start out in a smaller capability and then you can grow it as your needs grow. And it's incredibly modular, snaps together, and you can do the same with the IT stack as well. So as your needs increase, you can drop essentially our reference architecture and scale your cloud capability as well. All right, so we're looking through the latest survey result from OpenStack and of about 500 deployments that they had that took the survey, about half of those were on-premise. You know, what do you think about kind of that on-premise business? How does that fit with what you guys are doing? So we have approximately 700 customers that we have built relationships with primarily in the data center business. And we have very good relationships with those customers and they trust us with running a lot of their data center business. But what we're finding is there's still a bridge to be crossed regarding trusting of data and then trusting of applications. We believe we're in a really good spot because we've built that trust to seize that. But we're also seeing from people that, hey, from a data sovereignty standpoint, from a control standpoint, I need to have it on-premise. But to be honest with you, what we're also seeing is they're struggling, standing up an open solution on-premise. So that's where we can come in and say, hey, use our reference architecture, put it on-prem or put it in one of our partners' prem and or put it within our facility and you can have complete control of it as well. Yeah, of course, one of the biggest challenges enterprise have is they've got legacy. There's always, I want to try something new, but I have applications that I can't get rid of. They wish they could all be Facebook and say, I've got five applications that I just want to build at, web scale, what are you seeing in that space and what can be done to help customers bridge that legacy to the new and drag it along into as long as they have to? So I know we just talked about on-prem, but also really interesting in this is a lot of companies have taken the step to go co-location and they've done that over the past 10 years, rather than building their data center, they've essentially outsourced that. Well, that's only gotten them out of having to run the physical data center, but what we've been able to do is with our enterprise cloud solution, we can do a direct connection between what they may have legacy and Colo in our facilities and the data centers that we operate in Phoenix, Ohio, New Jersey and Singapore. We can do a direct connect, high speed connection, low latency at zero cost for Ingress and Egress between what they may have in their legacy area within our co-location and then our enterprise cloud. There's been a big scramble with some of the big cloud providers as to getting that global reach. Are you guys looking to compete everywhere or through partnerships? So partnerships are key. We've been very fortunate in that we recently announced a very big deal with CenturyLink. CenturyLink technology services has adopted our platform for delivering their services and what that's going to do is allow us to leverage their locations and their data center reach as well as ours to provide basically the solutions of our platform through multiple channels. Aaron, talk about the OpenStainers because obviously I took a tour of your facility, I was down in Phoenix and swung by so I met with George and the team and it's impressive, right? So it's impressive to see just the quality and just the success there. But now you're talking about OpenStainers, there's no software, right? So software defined data center is obviously the hot trend. So talk about what that means for you guys. How do you vector into that open source community? Obviously you guys were at Open Compute, that is a hardware version of the Homebrew Computer Club kind of in this modern DevOps era. So how do you, how do people connect that dots? What does that mean? You know, Open Compute, I get that, but how does that relate to OpenStack? So a little bit of history. We have had, we had a traditional cloud product that we founded about four years ago and we put it on brand name hardware. I won't use the brand name, but what we found is that it worked really well and it did in a function what we needed to do but to scale it was incredibly expensive and just not only from a hardware standpoint but from a software license standpoint. So about a year and a half ago, we set out to find a better, faster, cheaper way and really where that drove us was from an open standards on the hardware side to Open Compute and then from the software side to OpenStack. The combination of the two are very powerful and as far as I know, I believe we were the first to actually combine the two into a complete solution. But of course with that, we had the joy of doing engineering to make that happen as well. But a very compelling solution of, we now have a roadmap and open standards for our hardware solution and we can publish that and share that with all of our prospects and customers and we have the same on the software side. So talk about the footprint. I mean just for the folks out there, just give them some order and magnitude scale of IO because a lot of people can relate to Google Cloud, other of these other cloud as your others. What footprint are you guys dealing with from a physical standpoint with IO? So we operate over two million square feet of data center and when I say two million square feet of data center, it's active data center. Typically our customers are enterprise-sized customers. We also have done some cloud business in the past, like I've said, but we're also pushing very quickly into providing the enterprise cloud within our facilities and also providing that as a product to our customers. So we have defined our platform by providing the service for over 700 customers, two million square feet and we've taken those products and created the platform and made them available for enterprises. So you're at scale with Amazon, Google, Rackspace and Azure. So talk about the issues that are important. Scale, check, security. Talk about security. So from a security standpoint, some of the largest enterprises in the world currently trust us with running their solutions. So the company defines it. You don't have any, we have security capabilities but the company supplies that. So we provide the physical security and then a lot of our customers also rely on us for networking. So we actually provide the logical security for that as well. But from a cloud standpoint, where we really stand out security-wise is, is you have the ability to have your workload, your legacy workload in our co-location facility and you have the ability to elastically expand into our enterprise cloud over a local area network. Nothing goes to the outside. So one of the things is lock-in. People worry about lock-in. How do you address that? You say, hey, we're not no lock-in. How do you address that one issue? So in our cloud reference architecture, there is zero lock-in. It's open compute hardware of which we can procure from any ODM vendor that wants to go to that standard. And then we have open stack on the software side. We don't believe that there's a place in the future for brand names and the premium to go with them for most of the compute needs. There will always be specific compute needs of HBC and things of that nature, specific applications. But for the majority of workloads, for the majority of applications, a complete open solution will work. All right, so I'm the CIO. I'm sitting with all my staff and I say, hey, yeah, I've heard good things about you guys, a huge footprint. You're like one of the big dogs with Amazon, Google and whatnot on the square footage size. I want to go open stack and we're putting our plans together. What do you guys offer? What do you say to them? So we offer them a public cloud solution. So run on our facilities supported by our team. But really where our differentiator is is in a hybrid type solution. So with a hybrid type solution, leveraging our data centers, leveraging our existing networks, which are massive as well across our facilities, you have the ability to determine what you want control over, what you want to run and what you will allow to go out to the cloud. So the combination of those two and the ability to do it without any tolls across that connection and a very low latency connection as well. One of the things that we're seeing is some people have attempted to move workloads to the cloud in the past. So some of their applications, servers, some of their applications, but they couldn't move the database due to data sovereignty rules or whatever it may be. And the latency between the application and the database actually did not make it possible. It failed. We're seeing some of those folks come back in a low latency, high speed connection between the cloud and them. And how do enterprises evaluate the cloud? So obviously on-prem is pretty much a groove swing. There's certainly transformations going on in terms of footprint, power and cooling, a lot of always, there's always a gun to the head of the facilities managers on do more, reduce the power footprint. So I see that clearly as a huge opportunity for you guys just right out of the gate, low hanging fruit. Now the cloud is a little bit different. So there's still evolving. It's growing like crazy. It's clearly going to be a trend that's going to power decades of innovation. What do you guys see as the evaluation point in the RFPs? You're seeing it as, what are the table stakes? What are some of the key evaluation points? One is scale, because really with cloud, the ability to scale and go elastic is just a big use case for cloud. Another is security. So how can I trust you with it? Yep, another one that is ease of ingress and egress of data. So early adopters have found that it was very easy to put data into the cloud, but it was very difficult to bring data out of the cloud. And I think the other is interoperability and the openness, the ability to move my, to move what I have in the cloud from one provider to another provider with minimal effort. Okay, so for the future of OpenStack and IO, what are you guys providing? What's the roadmap look like? Can you share any specific high level data? I would say really the differentiator of IO is the ability to provide the same solution in our facility, in our partner's facilities, and your facility. So you will know what our reference architecture is. We're completely transparent with that. Unlike some providers, we allow you to come into the data center and with our OS software, we allow you to view every layer of the data center all the way from the utility to the UPS to the PDU to the server and then actual data from the server real time. So the transparency there and the ability for you to deploy that where you want it, when you want it, and in the volume that you may want it based upon your needs. You guys basically data center as a service on-demand data center, we already want to call it pretty compelling. So I want to ask you, what is the most exciting thing that you've seen come out of some of the work you've done with your customers that you can talk about? The most exciting thing that I've seen. I would say a complete fundamental shift in the way that some of our large enterprise customers are procuring compute. I'm not even using the word IT there, I'm using compute. So gone are the days of being called upon by six salesmen for hardware, software, that it's how they are procuring the ability to compute with a single SKU. That's game-changing, it's a paradigm shift. And you also got the ICO file and then you really are in a quiet place, you can't talk about that. What do you think about the future of OpenStack? Give me the final word, share with the folks out there, why is OpenStack so important and what do you see that going? I see the amount of momentum behind OpenStack is unrivaled by anything else out there that I see today. The only thing that even comes close may be the momentum that is in on the open compute on the hardware side. But from a software side, I really don't see anything else. I see the people, the big contributors to OpenStack coming together, coming together almost like a David and Goliath story. It's hard to even think that it's a David and Goliath story, but coming together to produce something that's better than what is available commercially and just continuing that forward. Yeah, the old expression, big iron was an expression that came out of really IBM's heritage, but that was being in the main frame kind of drove that 50 years ago. But now it's like, we always use the metaphors. The cloud is the engine and everyone's building their own hot rods. And if you look at IT, the cloud is powering mobile, it's powering big data, it's powering legacy apps, different workloads. So really the engines are basically going to be based on what people are driving. I mean, you see that same trend? Most definitely, most definitely. And you know, the coolest applications that we're seeing coming to the cloud are ones that just make sense. For example, one of our customers hires a significant number of summer interns and they do VDI. And previously they would go out and procure the hardware, install the hardware so that the VDI could support the summer interns. Now they're shifting that workload to the cloud. That's not a perfect application. I don't know what is. Well, I really appreciate you guys coming on. I know you're in a quiet period, Aaron Peterson. You guys have the best domain name in the history of IT and cloud.io.com. Can't miss that one. Certainly it doesn't help you on the search engine optimization. I mean, it's just two letters, iio.com. There's a whole story on that. Go to YouTube, George Schlesman goes into detail on how they got that domain. iio.com, great. You were always commenting on it because I was like getting, it's so hard to do. Thanks for coming on theCUBE. Really appreciate it. This is theCUBE. We'll be right back with our next guest after the short break.