 Live inside theCUBE, SiliconANGLE's exclusive coverage of OpenStack here in Portland. I'm John Furrier, join with my co-host Jeff Frick. And our next guest is Josh Bikenti, who's the CTO and founder of Piston Cloud. Welcome to theCUBE. Thank you very much. So tell us first about what you guys do as a business, then we'll jump into the OpenStack conversation. Sure, so we take OpenStack and we deliver a software product for use in building enterprise private clouds. And is that on bare metal or is that service? Everything from bare metal up. So we don't sell hardware, we sell just software, but we include hypervisor, we include virtual storage, we include virtual networking, orchestration, installer, you name it. Well, our tagline is typically cloud in 10 minutes. So, Sean Douglas from Service Mesh was giving you guys a lot of compliments when I asked him the question what people should do to get involved in the community. He said, you know, contact some of the thought leaders and mentioned Piston Cloud as one of them. What is happening with OpenStack? Share with the folks out there why OpenStack's so hot right now and why is all the excitement this year of all the other events. Sure, so I think Sean maybe gives me a little too much credit for having perspective in the sense that I've been doing OpenStack, I think longer than anyone. The first release of OpenStack source code was on my blog in 2010. So I have no perspective at all. Like if you ask me about the grizzly release, it's a blur, it's like four years of unbelievable growth. I think what's sort of fitting this time around is that the summit's actually in the Oregon Convention Center and this is where we announced OpenStack to the world. This was where the first public announcement was made at Osccon that summer. So it's kind of nice to come back, come full circle and be in this building where originally we had a handful of folks and now we have what this account this week is 3000 which is just insane. And they're expecting more the next event. Yeah. What's the makeup of the community right now? Obviously it's started with hardcore developers, you got a lot of suits here, got some business here, obviously the big whales are here, IBM, HP, VMware, other big public companies are here. But you still get the startups emerging in and you get the developer community. What's that dynamic like? It's really fun to be honest. So I sit on the OpenStack Foundation board, there are 24 board members and it's a really interesting mix because we have, we have small companies, startups, we have service providers, we have systems integrators, we have, as you said, the IBMs, the Intel's, the HP's and we managed to get together for these 10 hour meetings and get a lot of work done. It's really, everyone in that room is there because we believe in kind of the vision of OpenStack and how it can change what happens in the data center and we're all focused on that. Hey, we were talking about the cloud wars earlier, how the fight for the hypervisor and then the game elevated around the conversation around this idea for developers and building clouds which is a bigger vision than just spinning up virtualization in the cloud. And then now the commentary here, obviously we're at OpenStack, it's OpenStack One and it's the operating system of cloud builders. That being said, what is the cloud war situation right now? Obviously Amazon is doing very, very well. Are we in multiple cloud environments? Is it going to be that every company has multiple clouds and cloud brokering is the topic we've been kicking around? What's your take on where we are right now? Given all this, I mean, look at just Microsoft, we were talking about Microsoft just in the last segment and like, where are they? So yeah, Microsoft, you got VM, you got AWS. Well then there's the Google, the Google keeps sprinkling in but you would expect at some point in time the Google volume would increase as well. But there's no one size fits all cloud so there is conversations around, hey, how do you put a frame around this? How do you get customers getting value out of that? So what's your take on all this? It's a great question. There are folks who are very much that everything will be public cloud eventually. They believe that we all converge in public cloud. I'm not of that camp. I look at the internet as a great example and it's okay, well the internet became incredibly powerful because we have private networks and public networks built out of the same protocols. And so if you look at OpenStack as being sort of two communities in one, we have the private cloud vendors and we have the public cloud operators. Certainly in the public cloud, Amazon is the big incumbent. And I think the cloud wars as you put them are really between Amazon and the OpenStack clouds at this point and we see rack space and HP and IBM and others as being the new entrance in that market. I think in the long term this looks very much like the Linux ecosystem in the private space in that we will have a few pure plays that will have a handful of also rands. Every market eventually becomes a two horse race. So VMware is definitely one of those horses and the other one's clearly OpenStack but we're not sure which OpenStack. Set of steak knives for, and coffee for the winners as they say, so okay let's go back. So I've been following the whole OpenStack thing in rack space pre-OpenStack, prehistoric OpenStack as we called it earlier. And I was critical, I mean I'm a big fan of OpenStack, I always have been but I was really critical on the blog about some of the early days of OpenStack. It felt like a pool party where everyone jumping in for a splash for a marketing hype. Hey, Amazon's kicking ass, they have no cloud strategy. Say OpenStack and we're all kind of kumbaya and those industry kind of partnerships become like a Barney deal, everyone loves each other but nothing's happening. So that was my statement, I was afraid of that. You saw some signs, but it changed quickly. It did, a lot of people were afraid that OpenStack was more marketing hype than actual solid code. And I never had that fear because I was inside the code, right? So we saw, when we launched OpenStack Nova very shortly after we started writing it, it was about 6,000 lines of code. If you look at OpenStack today it's more than one and a quarter million. So that's an enormous rate of growth. Actually, Olo figures OpenStack is about 200 man years worth of effort. And the thing that happened is the companies that joined the community actually put in the work. Yeah, their marketing departments got really excited. There were a lot of press releases, there were a lot of announcements. We do a lot of speaking engagements. We're on video interviews like this. But folks are still doing the work. They're still moving the ball forward. What was the tipping point? Was it just the plowing the fields with good content, good code? The contributors themselves? Was it network virtualization? SDN seemed to kick things up a notch. We're validating from the big will. Where, I mean, was there a flashpoint? So here's, I've been working on interop for the last few months in the context of the board and I'm leading a panel later this week and I was quoted in the press last week talking about it. When we launched OpenStack, the NASA team, my team and the Rackspace team were 100% incompatible. We had zero interop in OpenStack. Even though, you know, so at that moment in time it was as much marketing hype it has ever been. And we've been converging that since then. I'd say we're 80% there. There was some point around the Diablo release where we actually got the auth piece working. You know, where literally you could use one set of credentials and you could use all the different OpenStack services. And that was the point where people said, okay, well now it starts to feel like it works. It feels real. They can put their hands on it. Yeah. Wasn't that like inter-fighting? That's good governance though. That's, you know, you guys, I gotta say, one thing that props the OpenStack community was they really moved quickly around solid governance. And, you know- Well, we got a ton of criticism for that in the first year. I mean, our original governance structure was two NASA people, two Rackspace people go. Yeah. Right? It was really more of the benevolent dictator model than we have evolved away from. And then we tried this thing called the Project Oversight Committee. We renamed it to the Project Policy Board. We changed the election structure for that two or three times and eventually we buckled down and put the nonprofit foundation together last year. We've really rapidly iterated. We've probably iterated on our governance faster than we iterated on code. And it's worked out really well. Yeah, that's good investment though. Yeah, it's interesting. So now that you've gotten here and it's exciting and it's big and it's real, what's the next challenge from the foundation's point of view that's kind of the next summit? Where's, where are the next priorities over the next time? That's a great, great question. You know, the original goal with OpenStack really was to reach the state of the art. You know, to say, hey, if we're going to call this cloud and by cloud we really mean the software defined data center or software defined everything, we've got to have storage that's best of breed. We've got to have virtual networking that's best of breed and we have to have our compute layer as best of breed. We've got that done now. You know, we've got Cinder matured, we've got Quantum matured, we've got Nova matured. Now we have this amazing challenge to, as a community, roll the ball forward. Where do we go for innovation? The last summit I gave a speech on ParaCloud, sort of talking about this idea of how does OpenStack become the perfect platform for every platform as a service offering? So whether that's OpenShift or Cloud Foundry or Chemologic or CloudBlease or JCloud or you have it, how are we absolutely the best place? What can we do in the infrastructure layer to make the application developers more happy? It's not about the raw infrastructure anymore. Now it's about those value-added services. Josh, I'm getting some tweets from some folks who follow us who aren't as geeky as us in the OpenStack world. Asked you to provide a OpenStack 101. Yeah, they're hearing Diablo, Novel, these kind of buzz words. So just quickly, just give an OpenStack 101 for the audience. Sure. What do they need to know? I can do it in five numbers, okay? OpenStack is one community around two kinds of clouds. That would be public clouds and private clouds. So in your own data center or somebody else's data center. It has three kinds of interfaces. So you've got a command line interface, you've got a web dashboard, and you've got a programmatic API. All three of those work in the same way and they service the same things. Now the fourth thing is you've got four kinds of resources exposed through those interfaces. So you've got two kinds of storage. That would be new object storage and old block storage. You've got virtual networks and you've got virtual compute resources. Those are your four resource pools. And finally you have five actors. This is the most confusing part, okay? You have vendors. That would be folks like this. That's confusing. Who sell clouds or cloud software. You have operators, the folks that run clouds, either again, private or public. You have auditors, people forget about these guys. They're the ones who have to look under the covers, but they're actually not delivering value. You have users, cloud users, meaning they connect to those APIs or the command line tools. And then you have end users. That's somebody who actually uses an application that's hosted on a cloud and they may not even know it's hosted on a cloud, right? So if you pull your phone out of your pocket and you connect to any app on your phone, you're an end user of cloud, whether you realize it or not. So that's OpenStack in five numbers. And then let me follow, then what was either the event that created the desire to create OpenStack or what was kind of the mission, the really high level mission when you guys got together with, however many people you were five years ago and here in lovely Portland. So I was running a team at NASA and we had a mandate to do something about NASA's data and how they were building applications. And it was a very broadly defined mandate because it was Skunkworks' unfunded project. And we originally set out to build a platform as a service that the agency could use. I mean, was it because their existing thing wasn't working well? Is it because they saw a new age of types of data or amount of applications? I mean, NASA had at least 300 data centers, didn't even know how to count them. Thousands of applications written in dozens of languages. So we had security concerns, we had cost concerns, we had legacy concerns. And really NASA's mandate is to collaborate, but we can't let people at our infrastructure because it's also one of the most secure government agencies. And so we were trying to build something on the inside of NASA that looked like things folks could run on the outside. That was the path to open source. Was to say, if we build this as an open source technology, we can use it. Other people can use copies of it and we don't have to give them tunnels through our firewall every day. Which sounds ridiculously forward thinking for a big government agency to actually say here with the crown jewels and some of the most secretive stuff I'm sure we have, as a government, let's use the open source model to drive the innovation to clean up the spaghetti nightmare. So if you look at my team that built us at NASA, none of us who were at NASA before this project and none of us are still there. So it was a very rare window in time. We had inspired leadership who said you can go do this, figure something out. And Rackspace was a motivated party to enable another group. Absolutely. Rackspace. Because they didn't want to take it on their own. I remember talking to Lou and Jim, Lou Mormon and Jim Curry about the time. They were like, hey, we're Rackspace. We have our own agenda. We want to build a company. They had bought cloud sites and they're trying to cobble that together. But they needed someone to step up. And that was really good timing. The meeting when we first met with Lou and Jim and the Rackspace team, they brought about 17 people out to NASA Ames. We spent a full day together. It was probably the weirdest experience of my life, to be honest. It's a lot like meeting a long lost twin. We sat down. Explain the meeting. So we had a couple of hours blocked off for them to tell us what they had built and a couple of hours blocked off so we could go over what we had built. And it didn't work that way at all. They got up on the whiteboard and they said, so we're writing in Python. And we're like, oh, so are we? And we're like, well, we're using Twisted. So are we. Oh, so we use these brand of switches. Oh, so do we. And it was literally like every single technical decision we had made was identical and we'd never met these people. We'd never talked to them. We didn't have common friends. We didn't have common backgrounds. And I'm like, well, you just wrote exactly what we were gonna write next year when we had time. And they had exactly the same experience for us. And after it was like, okay, well, obviously this is happening. If it hadn't been that perfect of a fit, it wouldn't have happened because this was six weeks before the first summit. We had no time at all. Yeah, so good time. So you had a shared mission, shared vision, shared execution, and how the personalities, you got the cowboys from Texas, you know, they're all cool guys down there at Rackspace. But what was the personalities like? It was, I would say everyone other than the civil servants had a blast. The civil servants had a lot of trouble because we were all breaking rules, right? Like who's allowed to pay for dinner when you take 27 people for dinner, right? Every single person had to pay for their own dinner that first night. That's good biz dev. Great story. And a lot's happened since then. So you talked about the five things for the one-on-one for OpenStack. What are you things going to happen as this goes forward? Because now you have a framework that's quite frankly in the public markets, in the CIO space or in the enterprise, it's like a warm blanket for CIOs because they need some help around feeling comfortable around their issues, which is like SLA, I got all this data protection issues, like I have all applications support, one-higher developers, and they're also going to a scale out open source model. So that is scary, it's not rack and stack, it's like I'm spreading servers all over the place and cloud is just another set of servers. So they don't know how to do that. They're scared. And Amazon isn't like the most comforting. Right, well any single vendor is not the most comforting. You know if you live in the world of Microsoft for a couple of decades, you start to realize any one vendor controlling that much of the core part of your business is a little scary. The thing that I'm looking forward to the most is open stack becoming very boring, to be honest, because what we have done is unbelievably revolutionary but what it does is it allows the application developers to do amazing things on top. When you take those constraints away, so I'll give you an example, when they shipped the Apple One, Wozniak and Jobs came out with this computer and for the first time ever, you could be a software engineer without being a hardware engineer. You could write software without understanding computer chips and this is when I started programming. I was six years old, I had an Apple II and you didn't have to understand hardware. I'm not a hardware geek. I'm actually not good with hardware. So open stack has now done that to distributed systems. Application developers can now write software that can deal with a million concurrent users without understanding anything about scale. Or ops for that matter. Or ops, yeah, don't touch the box. In fact, don't think about the box. Don't think about the filer, don't think about the network. Just think about the application and the end user. And so now I'm really excited that we get to just sort of disappear. The open stack community is really going to become a real one. What is the most disruptive thing that you think open stack is doing right now in the marketplace? What are we not disruptive? No, what's the most, what's the stack rank of top three? Top three, I would say storage is terrified right now, because they have been the bastion of proprietary hardware even more so than networking. Networking woke up to this a couple of years ago and said, hey guys, networking is going into the software. Networking has to match what happened with virtualization. Our hardware is going to become commodity. Let's become software companies, right? SDN was born out of that. That was a giant disruption. The major networking vendors have adjusted pretty well to that and they're moving now forward with their SDN offerings. The same thing's happening to storage. It's a giant bastion of proprietary hardware. It's being radically disrupted. If you go downstairs to the exhibit hall, Ceph and SolidFire and Violin and Nimble and all of these new storage companies with new thinking where their value is in the software. Well you have NetApp just running these two proposals and some code. Absolutely. They've brought some code to the table. They've been there from five years. EMC's brought some code to the table. You know, yeah, it's pretty interesting. EMC's brought code to the table? Yeah. Mostly through partners and channels but they joined the OpenStack Foundation officially. Last week? No, no, actually. About the time of the last summit. Had a dig on EMC. We love EMC, by the way. Well, Sean was on the show a minute ago. He's no longer worse for EMC right now. EMC's no longer involved in OpenStack. I was gonna say, what about on the application side? So you've seen a lot of stuff. What's one of your favorite application stories that's been enabled by this completely change? I would take one of our customers. So just as an easy, off the top of my head because I talk to them all the time. So Radio Free Asia is funded by, it's a non-profit funded primarily by the U.S. State Department and they operate internet services in Southeast Asian countries that have restrictive firewalls. Right, so just think of dictatorships where you can't get access to media, you can't get access to the internet. They run, you know, tour nodes and Wi-Fi hotspots and local broadcast, local language broadcast. What was it? Did it come from what it sounds like when you say the name? Was it originally just the super giant antenna on the ship offshore? Literally, literally, yeah, free radio for these countries with a democratic message. Now it's incredibly dangerous to be on the ground in one of these countries and be associated with Radio Free Asia, right? Like we're thinking, you know, rubber hoses. And so what they were looking for for infrastructure was truly lights out software defined data center which we were able to supply them. So there are no on the ground people at all. In fact, there's nobody on the ground in the country let alone in the data center. And so we support this remotely and we provide free internet access and local language broadcasting. I mean, when infrastructure is not something you have to think about anymore what the application developers can do gets really exciting. It gets really revolutionary. That's one off the top of my head. Yeah, that's great. That's great. Makes exciting times. Yeah, very much so. So final question is for the CIOs out there and people who have that legacy who are looking to open stack. I mean, I think everyone else on the that's got a clean sheet of paper is pretty much on Amazon at this point until you guys can make it boring and completely. A lot of people are coming off Amazon right now. Really? That's an, well, they're talking tomorrow morning. I don't want to, I don't want to spoil their notes. You can describe what they're saying with why they're leaving. Yeah, so there are a lot of SaaS companies or web scale SaaS companies as they grow up. They've grown up on Amazon. Like Zynga. Not like Zynga. Zynga is a great example. There are other companies that are less gaming, more real businesses. And that's a scale where they can justify bare metal, right? So bare metal is cheaper than cloud at a scale of one rack. Well, not have to hire all the administrators and the management side. So managed private cloud is the first step and then running your own private cloud is the second step. But to be honest, cost is only one of those factors. The second factor is the customization, right? So you see modern applications more and more. You might have an in memory database where you're looking at, hey, can we get a half a terabyte of RAM as opposed to maybe 96 gigs? What about a lot of SSDs? What about some Fusion IO gear? What about like the mix of hardware in the data center to push a next gen web app? It's not something you can buy from Amazon. Yeah, and it's all there right now. Or if you can buy it, it's an unbelievable price. Unless you're Netflix, and then you get the Flash. Yeah, but Flash has changed the game in the data center for sure. Absolutely. And with open source. Yeah, do you know what you pay for Flash or SSDs at Amazon? I mean, yeah. They don't really offer them to everyone yet, but still. Sure. 20X of what would be, what you could do it for yourself. Fully loaded costs with data center and everything else. And you know, the customers we talk to are saving literally 8X from Oracle and or VMware licenses by going open source with scale out. Easily. In the data center. Easily. I mean, Oracle's getting hammered right now. So final question. Well, they bought Nimbula. You know, that's an open stack member. I'm sure they can do something with that. They're now joining the open stack foundation. We love Oracle. Purpose Build is making a comeback. Yes. And it is, actually. We were talking about that earlier. That's a whole other conversation. Final question because we get in tight on time. For the CIOs out there who are really seriously are excited and or tooling up or preparing to plan for a modern era infrastructure inside their premise and cloud. What advice would you give them about open stacks, prospects, and them in general on their side of their business? So the biggest challenge that they're facing, whether it's open stack or any other software defined data center, is really that the change is not about your software hardware. It's a cultural change. And the culture that is driving this forward is the DevOps culture, right? So this is bridging developers and operations in a really fundamental way. The challenge for the CIO is that they only own the operations side of that business and they've jealously guarded it. So this is about letting down the walls of their organization and becoming a partner with the other C-level executives. Particularly the marketing group who drives a lot of application development in a modern enterprise. Look at that cultural change. Look at how do you really change how your operators execute in the data center and not what you're deploying. Because frankly, open stack is certainly the best solution for the software defined data center. It's not the only one. And the success of your project has way more to do with the cultural barriers to that adoption than it does to the technology. Okay, Josh McKenzie, the founder of Piston Cloud CTO here inside theCUBE breaking it down. Great 101 on open stack. What you need to know and just some great commentary around this evolving and growing community around open source and scale up, scale out, open source. This is theCUBE. I'm John Furrier with Jeff Frickley. Right back with our next guest after this short break.