 From San Jose, in the heart of Silicon Valley, extracting the signal from the noise, it's theCUBE, covering OCP US Summit 2016, brought to you by OCP. Now your host, Jeff Brick and Stu Miniman. Hey, welcome back everybody. Jeff Brick here yesterday and today. This is really what powers the cloud. This is the hardware, the real computers that are at the data centers that are giving us the hyperscale performance, the hyperscale applications, the support of mobile and social and big data. And all the things we talk about at other shows, this is where the big iron is. We're excited to be here for our third year. Big shout out to Pica8, Micron and Melanox, our sponsors for this event. Appreciate your sponsorship and all our other sponsors for other shows. Without that sponsorship, we can't bring out the crew and the cameras and the gear and the people to come to the events and bring you the signal from the noise. So thanks again to the sponsors. So Stu Miniman, you're joining me. Stu, we had day two at the show. Got to spend a little more time walking the floor, talking to some people the evening events last night. What are your impressions? Yeah, so let me break down and give a little flavor of the show. First of all, San Jose, it's a nice event here at the convention center. Couple thousand people. We haven't gotten an official number, but my estimate, probably in the two to 3000 range, the keynote yesterday was full. I mean, you go in there, good to see keynotes. Of course, you've got the face people kicking it off. Intel's the big platinum sponsor. I mean, they put a significant amount of money in here on the show floor. They're a big spot in the middle and they've got two of their partners, Hewlett Packard Enterprise, and I think it's Hive is the other one that's kind of three big booths going across. And most of the companies that are in the booths are what we would consider kind of a component company. So these are the chip guys, people that make SSDs, people that make the cabling or the connectors in here. But I mentioned Hewlett Packard Enterprise, so there's somebody that puts the pieces together. There are the rack based solutions that pull it all together. And then even you had, I mean Goldman Sachs has a booth here and I went over and talked to the woman there and she's like, well, we're really excited about this. We're contributing to it and we want to share, which is a little surprising. You don't think of the big financial guys as ones that are going to share because isn't that a competitive differentiation? IT is something that really helps enable a lot of what they're doing, but you see many of those companies really involved in some of the open source now. So it's interesting though, you talk about the competitive advantage for Goldman Sachs, it's obvious, the technology supports their trading business and all their financial services business. But what about for the hardware companies too, if a hardware company is getting into open source hardware spec, how do they find the place to differentiate, be consistent with the spec, yet still have something to differentiate themselves within their own competitive sphere? Yeah, that's a great point, Jeff. So you look at Microsoft's here and Microsoft created a spec and they said, here's what we're doing. And of course, since this is what they designed and they donated it, over 90% of the servers that they have follow that spec. Now, Hewlett Packard Enterprise and Dell, of course are saying, well, hey, we want to keep the huge partnership that we have with Microsoft and we want to be a supplier to you and we also want to be able to sell Microsoft, Azure Stack and Hyper-V and everything else to customers on-prem. So they're scrambling to make sure that they can support OCP and the differentiation is tough, Jeff, because we understand if we follow spec and we go, it comes down to price, it comes down to support. There's not as many places, but everybody, like any other open source project is you start with a baseline and you start adding on your pieces, growing the core and hopefully have that virtuous cycle of making everything better to be able to make it there. So interesting pieces and overall, I know Facebook looks at it as helping their efficiencies and driving down costs over time. So there's that pull if you're a supplier saying, I need to be a part of this because this is a massive part of the wave of change that's coming, but what's that going to mean to margins? What's that going to do to overall profitability? The other piece we're seeing Stu, interesting to get your take is kind of this bifurcation on the data center thing. You know, Floyer wrote a long time ago, big piece on the super data centers, basically because of latency. It all is driven by latency and this constant issue of getting the data and the compute close together, whether which one you move to, which is always kind of application-specific. So you still have the large data center, but then you've got the little data centers kind of driven by IoT and military and other demands where I need a data center in the field or close to the field or close to my factory or close to my sensors. There's really an interesting kind of dynamic in play between kind of the separation of those two kind of buckets of data center capability, if you will. Yeah, because it's interesting that there's so many examples of what you're talking about there, Jeff, about kind of that bifurcation because OCP on the one side, we're really talking about disaggregating all the components. When you talk about networking, the hardware and software is pulling apart. Compute, some of the pieces are pulling apart. We've seen what Intel's trying to do there. But on the other hand, when it's going to be purchased, you're really talking about buying something at a rack. So it's not, it's got all these pieces, but the solution gets put together and when it comes to kind of mega data centers versus edge, things are changing a lot. So rather than just saying, okay, here's my silo, here's what I buy application by application, now I'm thinking about where I put things. So I might buy SAS, which means, I don't really care where it lives. Somebody else has taken care of that, whether that lives on Amazon, whether it lives in a service provider. Customers are buying from lots of service providers, but you mentioned IoT is a real example that we're going to need things at the edge. There is an interesting French startup here supporting OCP that has this box that they seal five blades in it and the outside container, they fill it mostly up with oil and create a vacuum and seal the whole thing up and you've got a power and networking cable, but otherwise you don't need heating or cooling, it doesn't need to be touched. So great for some of those weird government applications or environments where I don't have the right data center to be able to handle what I have. So as always with IT, there's a huge spectrum of what we see going on. Definitely those large data centers, not only kind of the big seven or eight out there, but the service providers are growing these environments, looking to gain efficiencies as they grow and OCP is one of those tools that are going to help them improve their operations and drive down costs through overall efficiency gains. In the last trend still, let's get your take on this deconstruction of what used to be done with an appliance, right? Where you had proprietary everything inside of a box and it was optimized to run a particular way. And now I think really, it seems really the cloud in like Azure cloud that needs to be able to support multiple things. You're seeing it in software to find networking where we're breaking apart different pieces of traditional boxes. So even in the hardware space, it's still software eating the world and software still continues to be really the place that offers the most opportunity for differentiation and those efficiencies that before you had to have in this contained system are now being supplanted by really fast networking, really fast compute and really fast store. So Leo Leung who currently works at Scality is going to come on the program. He wrote a good article talking about the software that's going to sit on top because absolutely it's one thing to say that I've standardized on the hardware and hopefully it's less expensive but if it doesn't change my operation model, if I don't have the tools to really be able to orchestrate everything there, that's where just the huge amount of cost is in the last segment yesterday when we talked to Wikibon's David Floyer, he said it's $2 trillion a year spent on enterprise IT operations and that's where we have a big opportunity. It's not to get rid of jobs, that's not the purpose but it's to help companies be more responsive to the business and move faster and therefore we need to allow really it's that automation that intelligence of the system to take over some of those mundane tasks or not even mundane tasks but be able to respond to certain alerts and take care of things that we don't need that human intervention. When we talked to Kevin Lee from Facebook, he talked about the artificial intelligence, that AI stuff. We're just talking orders of magnitude more processes that need to happen and we've got those faster processes or all those GPUs to be able to crank through all the data, figure out what needs and give back to the humans more useful, more complete information at the end of the day. Right. And then what about the control? I'm always thinking of all these systems, what is the plane of glass that these guys are looking at most of the time? Because it seems to be there's so many systems now that can integrate and work together. How much of that can we automate so it doesn't have to show up on their plane of glass? Well, how's that battle going to shake out? Because as you said, orchestration is a big thing. How do you manage all this stuff? So at the end of the day, you got to be looking at something at the top of the stack. Where does that go? Who's going to win that battle? Yeah, so one of the real challenges, Jeff, with Hybrid Cloud is it's not one thing, it's usually, as we said, it's here on premises, it's the public cloud and it's lots of SaaS applications and there really aren't great tools to pull them all together. How do I create a portal? What alerts do I get? What has to happen? I haven't heard, at least in the parts of the show that I've gotten to see, haven't heard a lot about who's pulling together those pieces. It's definitely something that, for example, at the OpenStack show, there's lots of talk about how those pieces get pulled together. We've looked at many of the cloud management platforms that are out there. There's been some recent acquisitions. The big guys are all making acquisitions and Cisco, Oracle, IBM all make pushes because they have serious plays in those environments. But we'll see. Looking at this show, it's impressive to see Google shows up, why they showed up and what their real motivations are is something we're all debating a little bit. Is it kind of talent acquisition? Is it so that they and Microsoft can kind of stand up here and be, puts a little pressure on Amazon to who today is, I guess, the least open or least publicly involved in open source. So, what happens with that? Because all these guys making major pushes, growing what they're doing in the cloud. So, show this gives a lot of visibility on that topic. Exciting time. So we got another full day lined up for everybody. We're excited. We've got Melanox coming on again. Thank you for sponsoring. Melanox, we've got Scality coming on. Deterra coming on. Equinix, few other surprises in store. So hang with us all day. We're in San Jose at Open Compute Project Summit 2016. Day two of wall-to-wall coverage. Stu Miniman, Jeff Frick. You're watching theCUBE. We'll be back with our next guest after this short break. Thanks for watching.