 Okay, welcome back. We're here live in Silicon Valley. This is theCUBE, our flagship program. We go out to the events, extract the signal and the noise. I'm John Furrier, the founder of SiliconANGLE. I'm Joe Michaels. Dave Vellante, again here at really one of the most important events happening right now in the world. And that is the Silicon Valley magic is happening at Open Compute Summit. We're here to dissect it, unpack it, and broadcast all the knowledge and data we possibly can. And then our next guest is Executive Director of Open Compute Foundation, Cole Crawford, formerly the co-founder of the OpenStack Foundation. Welcome to theCUBE. You're on stage with Dave on a panel. I got to ask you, I mean, I have a good vibe here, but explain to the folks out here, what's the magic of Open Compute? Why is Open Compute a packed house this year? It really has got some great momentum, not hyped up at all, you've got players here. Real tech athletes, real heavy hitters, not a lot of vendor influence, so to speak. People trying to change the world. What is your take on that? Explain to the world why that's happening. I think you can compare why we exist to why America exists. I like to get patriotic every now and again. Everyone wants to be free, right? Everyone wants to be free. No one wants to be locked in. We got rid of that when we founded America. And if you look at the four fundamental principles of open source, those exact same things apply in hardware as they do in software, right? The ability to look at the code. We've got Gerber files, we've got CAD files, right? The ability to change the code. You can modify this stuff with your CAD software. The ability to redistribute that code, right? The ability to become a solution provider and let other people do derivative works on that. And the ability to let them do that. So there's some just phenomenal freedoms that exist inside of Open Ecosystems. It hadn't been done well in hardware before and we think that we've nailed that approach. Yeah, not just free of charge, right? So we love sports now, just baseball, Apple Pie and Open Compute. So that's kind of the America dream. Let's stay on that thread for a second because obviously there's a lot of controversy going on around a lot of values certainly in San Francisco Bay Area right now. You got Google buses being thrown rocks at. You got Adventure Capitalists kind of went off the deep end recently talking about the 1%. But innovation comes from communities. I want to ask you specifically this notion of free market, you mentioned America. America is a free market. Some are saying we want to avoid the rigged markets where it's actually, everything's rigged in advance. We see that with a lot of our government policy. We tried to open up the local loop decades ago, got crushed by the telcos. So let's talk about freedom and free choice in free markets versus rigged. What are you guys doing specifically to keep the magic of the innovation from being rigged? Yeah, great question and I think reality television is sort of like guided reality television, right? It's kind of scripted and I think that a lot of foundations, a lot of open source movements, best intentions in the world sort of end up in a place with the traditional model of, hey, here's the platinum level for X amount of dollars and with that level comes this vote. Well that starts as your community grows, as the foundation grows, as that movement grows, that price increases and now all of a sudden you've got very wealthy companies contributing to these causes. The Open Compute Foundation has worked and was founded, the laws and the bylaws of our foundation are fundamentally different in that the most important and strategic position you can have in our ecosystem is a project lead position which is a community elected position and that's important to know because there's no pay to play model here and I think people are seeing that if you're going to be a relevant strategic player inside of this ecosystem, you're down in the weeds. You're getting your hands dirty and you're working on technology and I think as we maintain sort of that pace of innovation story, as long as it's a community first and pace of innovation story that we're telling I think that the community's going to be well served and we'll be able to keep that match. How unique is this in your experience when you compare it to OpenStack, one of your babies, Linux, we talk about that a little bit. And so I was on the foundation formation committee for the OpenStack and when Frank and team came to me and said, hey, we've got this other foundation, this Open Compute Foundation. Frank and I both had worked at a tier one OEM and he said, hey, we're going to do it a little bit different. I was intrigued and I think that the way that our foundation operates is unique in that there's five board members for the Open Compute Foundation and those five board members are hands off in terms of the technical direction. If you were involved with Open Compute when it first launched, you would have seen that everything was about the 21 inch rack. Open Rack was the key cornerstone of Open Compute, the power efficiency, all of that lined up very well with Facebook's vision of what that rack should look like. But you've heard today from Frank and his keynote about folks like Fidelity, folks like Goldman Sachs. 21 inches doesn't work for them. So we wanted to apply those same design principles and it was community led. This wasn't a board of directors saying, hey, this is the direction we're going to go and I'll start coding, get all the engineers and the hackers together to start coding. This was a real consumer of at scale technology saying, this is our pain, we need our own aspirin and we're going to manufacture it the way that we see and that'll work for us. We got a question from the Twitter sphere out there on crowdchat.net slash ocpsummitv for the handle. When do you look at the value of the pay to play board seat? When do you kind of give a little bit to get corporate sponsors? So you got to support yourself, you got to run conferences. How do you balance that? That's coming to question from folks who would love your answer by the way and said, hey, we love that, that is the future. How do you govern that and how do you balance that? So it's a simple answer. We run very lean. We rely on the community. Facebook dedicates an enormous amount of people in our tiered membership model. So we're moving to a tiered membership model soon. That program's not yet rolled out. But it's a reverse model where the more you give us in terms of community participation and intellectual property and time, the less you pay for a better trademark. So our gold membership is less expensive than our silver and our platinum membership is less expensive than our gold. So it's reward to play. It's reward to play, right? It's the complete shift, you know, it's the opposite shift of what's been the traditional. Just start a publishing company. That's Tech Journalism, it's a bubble right now. So just start an event. You got an event, you know, throw a blog out there. Next thing you know, you're reporting on the news. That's a good business model. No, all joking aside, that's a good answer. And people want to know that. And so the transparency is what? The blog, how are you guys communicating that? Yeah, so we're very transparent. You know, the board meetings are, the minutes of those board meetings are generally published. The incubation committee where sort of things get voted. Those are, anybody can attend over the phone. We've got a mailing list, you know, all the usual suspects in terms of technology that you can interact with the ecosystem. Blog, mailing list. We've hired a great community manager, Amber Grainer. She comes from Lonaro, another non-profit. So we, you know, we try and be as open and as transparent as we can be. And you see, you know, even Microsoft joining the foundation, they had released information prior to them announcing here, which is a huge marketing benefit, right? Cole, I got to ask you about the, as I'm on this whole 30th anniversary of the Apple Macintosh. I was at the event in Cupertino on Saturday. It was a really great event. Went down there, labor of love. I loved my generation of programmers. And it was just great to see the celebration. They're hardware geeks, man. They're totally, you know, old school hardware guys, but now we're in a modern era. So I got to ask you, if you just shoot the arrow forward, go back from the original Mac those days of homebrew computer club. You have open source, now evolution. You got Apache, you've got all these different communities, all with their own taste of flavor. It's almost like fraternity row for open source. This one does that, that's the, you know, different kind of mode and mojo, each mojo. But now, with open compute, you guys are, to me, what a homebrew club would look like in a modern era. Big innovation, access to the cloud, seems to be the holy grail. Computer's one of the data center. So, you know, so now we're changing the game. What do you look for? What kind of magic are you anticipating? Or you hope to see of the open compute summit, the kind of magic that disrupted the PC and created the PC revolution. Is it an Amazon clone, more than that? I mean, what are you hoping to see? Well, I think, you know, hardware is irrelevant without the workload that runs on it. So, if you've got, you know, if you're interested in cloud or if you're interested in big data, there's a ton of open source software that you can apply to open compute. We've got an entire program dedicated to making sure OpenStack and Hadoop and these, you know, these various open source software projects run really well on our open source hardware. And that's a, that's, you know, it's sort of a defining moment in enterprise IT and at scale IT because for the first time, you can actually go and individually select the components of robust open source technology that you want to deploy in your data centers to fit the workload that's important to you. You know, you heard Frank mention a mile wide and an inch deep versus being a mile deep and an inch wide. And if you take the workload and you work backwards, I think, you know, yeah, we are hackers, right? And we're working backwards, but I think that's the right model. I believe that if Linus Torvalds today was writing Linux, he'd say, hey, I've got this thing that I'm going to write and guess what? You can port it to any architecture you want because it's all open. Yeah, and that's different from the time then. They were trying to get around the suns and the HP UXs of the world, the licenses as George Slashman from Iowa was saying, he pays zero license. So there's different, different conditions you're saying. Exactly, it's a different playing field. So we've got a question from again, from the crowd chat from an audience member. Ask Cole the biggest lessons learned in his participation in other open source projects, lessons learned, applied here. I can only assume he's talking about OpenStack where you were a co-founder and other environments you've been involved. So what are your biggest lessons that you can apply here to open compute that you won't either, maybe there are mistakes you wish you could do over or successes you're going to amplify on, double down on what can you share your perspective? Yeah, I mean, lessons learned is interesting. I think everything is a lesson learned and I want to go out and for the record say, we love OpenStack. We love the OpenStack Foundation. We love the Linux Foundation. This is a heterogeneous open source community that couldn't function in silos. So we have to be working together. I think personally for me, I want to be a champion for the consumer of technology. Again, you saw a slide earlier about a bottleneck between technology and the consumption of technology. And I think that if we remove those types of barriers, we can get to the type of infrastructure that solves cancer. We can get to the type of infrastructure that is processing 100 zettabytes of data in a year and coming out with relevant big data. Big data by itself is just data, right? I said that earlier. Relevant big data is what's important to us and I think the type of scale infrastructure that we're going to need, we probably don't even have that yet today, right? We're going to need disaggregation. We're going to need SOCs. We're going to need multi-architecture platforms to run a number of different specific workloads on. And I think as long as you put the consumer first, as long as we're being champions for the consumer, then we're going to be in a good spot. So let's talk about disaggregation. It was a lot of talk last year at this event about disaggregation from a consumer's perspective. Why is disaggregation important? And then the other observation, it's not really criticism, but the other observation about OCP is it's designed for mega-scale only. It doesn't scale down. I wonder if you could address that as well. But start with disaggregation. Why is that important for the consumer? So Patten, General Patten once said, we're getting patriotic again. He once said, fixed fortifications are a monument to the stupidity of man. And I think through disaggregation, you end up being able to, yeah, you've got your fixed data center, right? But you also have this eventually consistent ecosystem or this eventually consistent data center that with the future backplains and the future MPLS circuits, the layer two data center connects that are going to be coming out in the next five, 10 years, we're going to see bandwidth that we just don't see today. You're going to see replication happening much faster. We're going to need to, right? With the amount of data that we're processing at the distributed object stores and the distributed storage that we're having to deal with, you're going to see eventual consistency play a major role in the future data center. And I think this Ag is important because no longer are you going to be, you know, you're going to be forced to rack capacity, not rack servers. Racking the days of racking servers are over if not, you know, a short lived reality. The ability to rack- Racking stack is a dead end, is essentially what you're saying. It's exactly right, right? You need to be thinking about racking racks, not racking servers. That's the new world that we're going to- Racking data centers. Exactly. Hey, we're going to put that data center, let's put on top of the other data center. I was telling to George Schleschman, I think the future is shipping data centers, not servers, that is the future. That's thinking it's more containerized and smaller. So what about, what about that? It's sort of a soft criticism, right? It's met from mega scale OCP, but not scaled down. Is that true? Is it a problem? You know, I don't know if it's true or a problem, right? I think we're at this place where the world is our oyster effectively, right? The data center, as we move into bigger and broader data centers, power efficiency is just massively important. And you know, the IOs of the world are sort of leading the charge. I mean, these are bleeding edge technologies that they're adopting, but you know, you look at open source and the ability for George to do a live demo on an iPhone, deploying OpenStack on OpenCompute. You know, I think that's the future. He said it, I agree. Yeah, and I think, you know, a lot of things you kind of were giving an answer about OpenStack earlier, like you love them, like almost like you feel like people are almost implying that OpenCompute is competing with OpenStack. I don't think I've ever said that. I don't see that. I see this as a symbiotic relationship of what it gets done in the engine. And certainly at the hardware level, as it gets shrinks and gets more performance and with software virtualization, network virtualization, we all know the benefits there. This is a boom for OpenStack. I mean, this is absolutely. And they're a boon for us, frankly, right? I mean, Graham Weston, the chairman of RackSpace said, he compared OpenStack and OpenCompute to peanut butter and jelly. So, you know, and I think that that relationship really does apply. At least in terms of, if you look at cloud, which is a major at scale workload for a number of managed service providers, a number of telecommunication companies, a number of, you know, big data companies, if you take all of the goodness of open source software and open source cloud that OpenStack brings, and you take all the goodness of the, sort of that open API at the hardware level that OpenCompute brings, you know, you heard George say it. There is not a single license fee that they pay to be able to deliver that experience to their customer. Well, we love cloud. And one of the things you'll see from SiliconANGLE Wikibon this year is a lot of cloud coverage. We don't really kind of toot our own horn about saying that we kick ass and cloud. We're pretty bad ass. We know a lot about cloud. So, I got to ask you, in the cloud conversation, all the suspects are there, but you don't see in the mainstream press yet, Facebook. So, they have a huge cloud. It's called their application, called Facebook. The SaaS is their application. And they're a big part of this environment. So, you know, speculation inside of what they will or may not do in the future relative to being an Amazon competitor, let's, to be another day. But I want to talk about their role here, right? Facebook's role with the Open Compute project. How's it evolving going forward? And two, the second part of the question is, your take on Amazon, because Amazon, in a way, has replicated, kind of, in their own way, build their own gear. They make their own, they source everything and build it in house, smaller footprints, the densities, tiny. They are their own version of Open Compute. In a way, the gold standard today. So, these are two massive market whales out there. And John, James Hamilton, you asked James Hamilton the question at Reinvent. What about OCP? Ah, we love OCP. But it doesn't quite do it for us. So, it's an interesting discussion. So, this is kind of converging together. You got, you know, the two whales out there. And Google's, you know, obviously sitting in the wings looking down, probably going to, you know, what's that old guy, I won't go there. But go ahead, talk about Facebook, their role, and then the Amazon, kind of, I guess. The horse is on the track. I mean, how do you look at all that? Yeah, so, you know, with regard to Facebook, I can't really comment. I've never been a Facebook employee. You know, obviously they're involved in the project, but no different than, no different than Goldman Sachs is involved, or Fedoli's involved. Obviously, they're big consumers of Open Compute. And like Rackspace, you know, I would compare Facebook to Rackspace in that they came up with this neat thing, right? Rackspace had OpenStack, in combination with NASA. And Facebook had Open Compute. So, altruistically, they did the same thing that Rackspace did. So, I would put them in sort of that, you know, altruistic creator. They're a leader, too. I mean, they're a leading and they're a catalyst, and right, okay. Definitely, they lead, and they, if you look at how much they're- They're by a boatload of machines. I just tweeted, they saved $1.2 billion this year. Jay was just on stage. That's on the Twitter sphere, on the crowd chat. So, I mean, obviously they're huge. Huge consumers, too, right? I mean, the Lulia data center is a massive Open Compute data center. So, if you look, again, if you're looking at the consumer first, Facebook being a consumer of technology, they want to get to the most, so there's a metric in data centers called PUE, Power Utilization Effectiveness, and they want to be the most efficient they can be, because that's how you get your better TCO. That's something where you do want to be 1.0. That's exactly right. You do want to be 1.0 there, and there's numbers published from Facebook of 1.07. I've seen numbers closer to 1.03, so they're unpublished. So, that is massively efficient hardware. And just on the comment about Amazon and Google, I'll just say this. We believe that anybody that thinks differentiating in hardware is a value proposition to themselves is doing it wrong. You know, I was talking about the Mac thing, go back to the gettings, it's so fresh in my brain, because this is such a parallel to me. Bill Atkinson said, we built the Mac for ourselves. In a way, Facebook is building Open Compute for themselves. So, this is why the magic is happening here, in my opinion. This is the industry saying, we want better, faster. So, do you agree with that statement, and can you give some other examples of people building there for themselves? Sure. Again, let's take Fidelity and Goldman Sachs, right? Building for themselves. We start out as a 21-inch rack standard. Goldman Sachs, and you look at the financial services industry in general, and these guys are bleeding edge adopters. Look at the Linux kernel. They were responsible for the real-time kernel. Like, these guys were the catalyst for creating the real-time kernel in Linux. That same thing exists here. They've got real estate close to Wall Street with fiber connections where millions of dollars are the difference between microseconds for trading, for high-frequency trading. And they say, we can't deploy 21-inch racks because we've got this co-located facility. The power requirements are so vastly different through what we operate today. We love the principles. We want the technology. We want to be able to build what we need to build for us, but we need to do it in this, with this constraint. Cole Crawford, the executive director of Open, Open Compute Foundation, formerly the co-founder of Open. Stack, great guest here, great innovator, great person to drive in the community with the team. His Twitter handle is Cole in the cloud, at Cole in the cloud. Thanks so much. Good luck. Congratulations on the fantastic event. You got lightning in a bottle here. Don't screw it up. No, seriously, have a great show. We love it. We're big proponents. They're building the future for themselves. This is the community of OCP, Open Compute Project. This is theCUBE. We'll be right back with our next guest after this short break.