 Live from the San Jose McKennery Convention Center, it's theCUBE at Open Compute Project U.S. Summit 2015. Okay, welcome back everyone. You are watching theCUBE. This is our flagship program. We go out to the events in extra discipline noise. We are here at the Open Compute Project Summit 2015. Hashtag OCP Summit 2015, join the conversation. I'm John Furrier with Jeff Frick. Our next guest is the president and chairman of the Open Compute Project, Frank Opsky. Welcome to theCUBE. Thank you. Go ahead and be here. We really want to say congratulations for your success. You know, you gave some perspective of where it's come from, Project Freedom. Did you expect there's going to be this big? I mean, this is now an absolute top of the first inning industry event now. Started out as a bunch of hackers, guys tinkering around, building some hardware, going open source, bringing two worlds together. And we covered the first Open Compute here and it was an awakening. Like, hey, these are worlds that are on a collision course. Let's bring open source to the table. Are you surprised? What's your personal feeling? What's your view on all this? Well, you know, I always had envisioned that we'd be able to drive change in the industry by applying open source principles to the hardware industry. But I had no idea that it would happen this quickly. Right? I mean, if you look at the industry norms that have been established over multiple decades to see them change as much as they've changed in such a short period of time, I think it's a really an endorsement of this community, of the voices that are in this room. And when they get galvanized and they say, look, this is what we want it to look like, right? I know my environment. You know, you've been underestimating my intelligence level as a supplier to me. Just listen to me, right? I know what my environment needs. I want it built this way. And I think that's really what has driven the change to happen and happen so quickly is that the people that are part of the community are so passionate about it. They volunteer their time. They volunteer their IP. You know, I think Microsoft is a great example. I mentioned it in the earlier keynote where I think everybody was skeptical last year of like, why is Microsoft joining an open source project? Right? What did they know? That was the beginning of the seed change of Microsoft. I remember we talked about that, like, they really donated, they laid down some goodies and that was not just a throw away. Fields, like some serious stuff that they laid down on the altar of open source prior to Balmer being kind of ousted. So like, that whole thing was revolving. What have they done since? Can you share more about Microsoft and some of the other big guys like HP? Yeah, so I mean, Microsoft has contributed a lot of time to the project. So we have people like Mark Shaw who is the project lead for the server project doing a fantastic job with that part of the community. Koushagra Vaid, Bill Lang, who's now on the board of directors at Open Compute. So the time investment, I mean, even the surface tablets that the engineering breakout units people were using, they were all contributed by Microsoft to be utilized during the summit. And then you look at the other side and the IP contributions that they've made, they've actually contributed more IP over the last 12 months than anybody else in the community. So, you know, what was a skepticism last year about like, why is Microsoft joining? Are they actually going to be an active contributor? Is this really just a PR thing? They have clearly come through in spades with the way they've contributed to this project. And we hope to see that pattern repeat itself with companies like HP or Cisco or Apple as we announced, Apple is a new member today as well. You know, I mean, I think it all starts with, you know, that collaboration. And a lot of times people come to the events and they see what they're about and then they join as members and then they start contributing their time and intellectual property. And I hope to see that pattern repeat itself with companies like Hewlett Packard, like Cisco and others. Talk a little bit about the juice that that provides inside the company, not from the open source side, but from the employer side. I mean, you were at Facebook, this was launched out of Facebook, but then now you're doing the open compute project. What does that do for the folks inside of Facebook, the management of those folks, where they spend their time and what kind of juice and feedback and energy does it provide that they weren't getting, not being part of an open source project? I mean, it's a source of pride when you can actually share your work openly with other colleagues and peers within the industry for an individual engineer who is used to the industry norms of building hardware, which was, you sign on as an employee and you can't even tell your family about what you're working on and then you launch it to the world, right? And it's like, look, we created the solution. We don't know what the problem is for the solution yet, but we created the solution because we really haven't talked to anybody about what we're building, but we're hoping it's going to be a fit somewhere for companies and for individuals within those companies that are able to openly share their work. It's just a source of personal pride, and it ups their game, right? Because when they start sharing in these engineering workshops and another engineer from another part of the community says, well, why did you do it that way? And did you think about doing it this way, right? It just creates this awesome cycle of friendly competition and it improves the designs. And so it's obviously a source of pride that they get to share what they've been working on, but then it's also upping everybody's game because you're sitting in a room with 30, 50, 100 other people who are just as skilled as you are saying, why'd you do it that way? Did you think about doing it a different way? So the question that everyone wants to know, we brought this up in the intro, this is more like a high level question. So I want to get your kind of like, just tease out the landscape. There's a unicorn in here, or maybe a few of them out there by BC students. A lot of BC's shocking in here right now. It's early on, this industry's really looking good. It's the beginning of large scale systems in the open. So huge opportunity. So smart money's in here. I see the guys walking around. You know who they are. You're at Facebook, so you know the deal. So the question that comes in from some of the folks in our community is, where's the next wave of power of virtualization slash containers? And is it going to create any economic value to entrepreneurs? Where's that value? Because with the software model, cloud virtualization was really boom, big part of the disruptive enabler. Now you got containers, which is really make software development really cool. And then as you move down to the hardware level, where's that next wave? Where do you see the next opportunities with white spaces for developers? Because you know, you're talking about software behind data center. It's the kernel. Like being at BSD Unix in the early days, if you will. But like, you know what I'm saying? So what's the, what's your perspective on that? I mean, there's no real right answer. It's more of what you're seeing anecdotally. Well, I mean, I think. So first of all, let me correct you. I'm not actually with Facebook anymore. Oh, you'd be, that's right. I left almost a year ago. But still very connected with Facebook. They're still a huge contributor to the project. You know, technically not with Facebook. Okay. I'm technically not an employee anymore. Still in the family. Still part of the family, yes. We're all talking. I think, and I mentioned this earlier, but in order to build any sort of software defined anything, trying to build it on top of a proprietary base is kind of like trying to build, you know, a tall building on a shifting sand. Because you don't really know. You don't have transparency of what's happening underneath. And typically as a developer, you're trying to actually work around some of the proprietary hooks. And so I think, you know, the big opportunity that Open Compute provides is to bring full transparency to that underlying base so that as a developer, whatever X is, software defined X, whatever X is, you at least can build it on top of something that you understand completely down to the BIOS and firmware level. What those unserved opportunities are, I mean, there's obviously a lot of them. But you know, one that I personally kind of think is cool is this idea of data center orchestration, right? And so a combination of really deeply understanding, more deeply than a traditional building management system understands what's happening at the node or the rack level, really getting a good understanding of what's happening environmentally, of how much power is that rack or how much power is that row consuming. And I've got this fixed provisioning model of power that maybe some of it is stranded. Being able to orchestrate and give people the analytical intelligence of, hey, you know, this particular Hadoop cluster is really only busy during these times of the day and my recommendation out of this engine is to do this with the physical layout of your data center, that, you know, to be able to reshape how people utilize the very expensive infrastructure that they have. Yeah, it's almost an adaptive fabric kind of model, software model, think about the container of a PC or a server, yeah. You have stuff inside the server, if you just take the shell and make it a data center, yeah, it's still subsystems, it's still network management, it's still software needs to coordinate resources, and if you look at linking and loading, all this stuff, that's an operating, these are operating system concepts. And I am sure that somebody can implement that in a completely proprietary way, so that if you have 100% from the rack all the way up through the top of the stack, if you have 100% XYZ vendors gear, they can make that work. In order to make that work in the real world, I think open source is actually an imperative, right? Because if you're not sharing what those interfaces are, if those interfaces aren't open, how are you ever going to actually orchestrate all of these things to work together? You know, one interesting thing I want to give you a perspective of, back in the old 80s, back when I was a young buck, you know, IBM, DEC controlled the network protocols, the stack was fully verticalized by vendors and the mainframe minis, and then, you know, all the side model came out, right? And then you got some standards, but it truly wasn't standard. You had TCPIP, kind of like, that was it, right? And then you could argue what got standardized. So that being said, that was kind of the model of open development back there, and that created a lot of wealth. Cisco, 3Com, you name it. I mean, the DNA off the open standards from those days is what we live today, Intel, Cisco. So now today, I want to ask you kind of, with that kind of in mind, open source, you have the purists on a software basis, red hat, cloudy hair for a dupe, those are purists. Now you have the open core model. So open source is taking a whole other trajectory. What is the new future model? If in the OSI model, there's no proprietary nature now, so like you had some big whales fighting for the stack positions, there's no proprietary nature in anything right now. So is the open core, is the purists, they're not necessarily mutually exclusive, some say, what's your take on all this? Because the innovation's coming. Software will be there. People do have to differentiate. So can you like just share your perspective of, is there a one-size-fits-all open source? Is there multiple versions of it? I mean, there has to be more room to grow. So there's a debate on what this governance issue is, open core versus pure, and you guys are in a unique position. I want to get your take on that. I guess my perspective is that standards committees have played a very important role in evolving the industry to the point that we got to. But as you mentioned, the whales were really the ones leading those standards efforts, and they weren't leading them out of the goodness of their hearts, they were leading them because they had a commercial opportunity if the standard got set this way versus this way, right? Then you flip over to an open source model and the purists would just rather say, look, you know, this is the way customers want it, so this is the way it needs to be built. And I don't think that what that means is that the entire stack just gets completely normalized and commoditized and there's no room for anybody to build anything proprietary on top. As a matter of fact, I think the reason suppliers engage in open compute is because the things that have historically been what they consider differentiators, you know, I've considered them gratuitous differentiation and they've held on to them for too long. Let's just normalize and commoditize that. That's a good point, let's talk about that. Differentiation is different than lock-in. So you can have differentiation proprietary, proprietary meaning unique to the company. Proprietary differentiation should be considered innovation. Yes, exactly. And gratuitous differentiation is the stuff that just gets in the way. Yes, exactly, yeah, yeah, that's a good, that's the way to parse it. And if you look at it that way you can identify dysfunction from function. So rip it out of the wall of the grass. Weed growth, so gratuitous differentiation versus what was the other one? You said just, oh, good differentiation is innovation. If you're differentiated in a way that matters, if it's relevant and it solves a need that nobody else can solve, then that should be yours and you should go make a ton of money on that. Yeah, and I think that the circumstances are now big enough where the opportunities are bigger. I think I personally believe it's going to be in pure open source, which is great. And then the open core gives people some comfort, maybe slower for integrations, like I saw on IBM. They have huge issues around slowness. They want it to go slower because they got to integrate a lot of legacy. It's not a green field opportunity. But if you're in a green field building stuff from scratch, you want the best. So it's interesting to watch this new era of open source develop because you have to have profits. If you don't have profits. Well, then one important part of your community just leaves. If there's no opportunity in it for suppliers, they just run for the hills. And so you have to have that. And so in many ways it helps suppliers, for example, Facebook and Intel talked about the innovation of the new chip that Intel introduced based on feedback from Facebook. Well, the thing that's interesting is it was feedback from Facebook, but it was also feedback from the open compute community that if you build something like this, we will adopt it. And so it makes what used to be speculative R&D investments, it makes them look a lot less speculative because you at least know after having attended these events that I've at least got a community of people. So if this one customer's forecast is wrong, I know that I can at least recoup my R&D investment by going to talk to all these other customers in the community. So as a supplier, one of the key values that they can get out of this is just having this massively scaled focus group that they can hear a galvanized voice from and say, oh, okay, well, that looked like a silly idea, but if they're all going to buy it, let's build it, right? Well, Frank, I know you've got time, you've got a hard stop, and we appreciate you coming on. Congratulations on all your success. We're watching open source. We believe that collective intelligence is the ethos of open source. The more people involved, the better the outcome. And now it's mainstream, so it's getting bigger. The pool of open source developers is so massive. And what you've done here with your team and now the community has been really impressive. Congratulations. And we think there's going to be a ton of more innovation in open source, not just the purists. I think you're going to see spectrums of innovation. So we don't want the gratuitous differentiation. That's bad. But the good differentiation is innovation. Thanks for coming on theCUBE. Live here in Silicon Valley, Open Compute Project Summit 2015 with the president and chairman, Frank. Thanks for coming on. We'll be right back into this short break.