 New York City for a special CUBE presentation for HP's Moonshot announcement where they're launching the big game changing product moonshot where it's changing the game in the data center, low power, high performance, powering cloud mobile and social. This is theCUBE. We're attracting the signal from the noise. I'm John Furrier, the founder of SiliconANGLE.com. I'm joined by my co-host. I'm Dave Vellante of wikibon.org and Regine Skillerin is at theCUBE in the house. She is the director of marketing for cloud at Intel. Regine, welcome to theCUBE. Hey, thank you so much for having me. Oh, you're welcome. Props, Intel and the Catbird seed and the HP Moonshot announcement, congratulations. That's great. Good to see you. We've been seeing some good innovations today. What do you have with you? Today, I've got the first production, the only production moonshot system that's... Today I've got the only production moonshot system that's shipping on the market today. We've been working for a couple years with HP and developing around the system and developing servers for the system. And we've been talking about it last year, but this product is now shipping. Shipping, as you heard this morning, it's shipping in Asma and Canada and the Americas and Canada and then it'll be going global next month. We've been seeding it to a number of customers jointly. Yes, you may. We've been seeding it to a number of customers, customers you've heard from this morning like Savas and Lee Swib and we're getting a lot of good feedback from our customers. Yeah, so you focused on cloud. The cloud guys must be excited about this announcement. John and I were talking earlier, when people see this, they say, oh, I need some and they have a sense as to where they're gonna put it. What are you seeing in your constituency base? Yeah, I've been following this segment for a number of years now and have been really doing a lot of research with our end customers, whether that be hosting providers or web service providers or even enterprise looking at different applications. And what we've really found is that there is a class of lightweight workloads that benefit from the lower power, the increased density and of course the performance for what this type of system can provide. And we see it working really well for things like static web or we talked a lot about analytics, batch offline analytics is a great use case. And of course, we've had per the hosting, the seeding I just talked about, we've been seeding with the hosting providers because we think it's a good dedicated hosting solution in the low end. Yeah, so analytics has come up a lot. So you see this as a big data platform, right? That's the hot term, hot word. And so what makes it so applicable for those types of workloads? Well, so what we've found is we've done a lot of workload testing. Intel has our own Hadoop distribution which we launched into the market. We've done a lot of testing plus a number of our key customers have deployed Hadoop frameworks and done a lot of analytics work. What we've found is when you do things like the offline and batch analytics, you don't need as much processing power, you need more IO and throughput. And that's when this type of system comes in play. We've also seen though that there's some nodes like the data nodes that hold vast amounts of storage and they might need a lighter weight CPU with it too. And then you've got your processing nodes that are called the data nodes and you might need more compute power. So what we see when we're looking at analytics, implementations and deployments and people who are then moving into real time dynamic insights gathering from their data, they're doing the ability to mix a range of solutions and having that broad spectrum from the lowest of power and the lowest of performance up to a more compute intensive system like Xeon. Putting those together and having software compatibility across both those environments is really what's gonna drive the best TCO for a big data or analytics type deployment. So it's critical that you have both. Yeah, so I think a lot of people were somewhat surprised that Adam was gonna be the first instantiation and didn't understand actually there's gonna be that cross compatibility. So that's big. And so talk about how those two coexist. How do you see those worlds sort of working together? Working together? Five years, let's say. Well, I think the systems are gonna evolve. The HP obviously has a broad spectrum of systems in their portfolio and although they haven't made any Xeon announcements around the moonshot system, they have talked about continuing to invest in their Adam lineup, which this is our first. And then this year, we're also gonna have one based on the Adam processor, which is the next generation until Adam silicon. This is based on our 22 nanometer technology manufacturing process. And what you're gonna see is increasing that performance and efficiency all with the software compatibility. And then in the future, HBN until we wanna continue to collaborate across the range of systems and really right size the hardware and the infrastructure get application optimized or workload optimized products out across the whole spectrum of key applications. So I7s or I5s, which in theory should be able to slot in. There are different, I mean HP, like I said, has not made announcements in this place, but what we're really looking at is saying whether it's for technical computing and being something like a Xeon 5, really needing a silicon architecture that's been designed for high throughput and performance computing, or whether you need Adam where we're gonna continue to take lower, lower power levels and increase the performance efficiency, you're gonna have the silicon you need across the range of platforms. Standing. So a lot of people talk about the four trends, the four big ones, cloud, mobile, social. Not big data. Big data, right? It's kind of the fourth. That's sort of the edge, if you will, of the enterprise. Is that the way to think about these platforms? Is really that's where they're gonna attack first? You mean in the cloud space first and then move in? Well, I just think right now, there's a lot of different, there's different needs based on if you're an enterprise, based on if you're a cloud service provider. When you're building out at hyperscale and you're managing tens of, not hundreds of thousands of servers, and especially companies that are service providers where their cost of sales is their data center, that's a very different environment than an enterprise. So it really makes sense that those type of classes, companies would be taking their workloads and saying, I'm going to find the optimal infrastructure, the optimal platform that's right sides for each of my workloads to drive out every bit of cost and power saving. Because they're able to make money off of their data centers as opposed to be a pure cost center. Yeah, this is how they increase their margin is by driving efficiency. So while the vast majority might be too socket today, it makes a lot of sense that now that we have these types of processors with the Intel Atom processor where we brought it on to the server roadmap, added critical data center features, things like the 64-bit software compatibility so you can move it into your environment without needing to port, as well as ACC. That's a huge feature right there. You can run it today. So you can buy it today and you can put it in your application environment today. And for those web workloads we talked about, front end static web or dedicated hosting or simple CDN, great applications to start testing on this platform. We'll talk about the impact to cloud. So you know a lot about the data center. We've had other events together where we talked about what's going on in the future of the data center and power and cooling is a top conversation as well as software manageability. All the stuff that you guys are used to Intel's not a stranger to the data center. But now with cloud you have this new dynamic called Amazon, which is totally leveling the playing field and forcing everyone's hand to say, hey, you know what? That is a viable option as a resource. But most enterprises we talked to don't have that comfort in managing Amazon. They're working hard to get there. So what does this fit into this cloud service provider? Because they're advanced from a service standpoint. They have diversity of services from multi-tenancy to other variety of things. How does this help them? How does this product, this direction of low power? You mentioned Hadoop as a use case. That's a good one, good one. Is there other use cases that you can see with the cloud that where enterprises can get to hyperscale either by retooling or going to the cloud? Well, yeah, so I think they can do both. We spend a lot of time with enterprises because we think that a private cloud can be built out very cost effectively in its scale. Maybe not the same scale as in Amazon, as you mentioned, because economies of scale and purchasing are always gonna be different. But a lot of the technologies that Intel created first that we delivered out to some of the world's largest cloud service providers, the global leaders, we've trickled them down into our product line. And what we wanna do is work with companies whether an enterprise is building out their environment to make sure they have the most efficient infrastructure. And this is where we wanna work with HP to position appropriately, may not, like you said, this may not be the platform moving into enterprise immediately today, but everything does move. And today this is based on the segments HP's been talking about more of a service provider move. But what we also want them to do is help service providers be more competitive as they create those cloud services. So the more efficient their infrastructure, the better the service and quality they can provide. Also we also Intel kind of separately from this, we spent a lot of time kind of trying to create technologies that enable service providers to differentiate their top technologies and provide additional value. And that's I think part of the push-pull too. Not just costs, but value added differentiation. Yeah, so that leads me to my next question. You talked about earlier about your own Hadoop distribution. We had Boy Davis on at Strada, right? He came down from San Francisco right afterwards to be on theCUBE. And so you're doing obviously, Intel's always done a lot in the ecosystem and software, but it seems like there's an increased emphasis there lately. What are you doing in this space from a software perspective? How important is software? I mean, obviously a software is important, but what role does it play in terms of the adoption of these types of new applications and use cases and what specifically is Intel doing to foster that? Software is absolutely critical. One of the concepts, and you probably heard Boyd talk about this last time, is looking at the data center as a system. And that means not just your compute, network and storage infrastructure, your standard resources, but looking at the orchestration layer and the facilities integrated. And one of the things Intel has been reshaping itself and has been making investments for a series of years. I mean, we've bought large software companies and we've been building software internally for many, many years and working with vendors to optimize. But what we're also doing is making sure that as we build out this orchestration layer, as we build out the solutions like what Boyd did with the Hadoop distribution, that we're making sure that it takes advantage of what's in the hardware so we can not only give a full reference stack that's gonna work well together and gonna be easy to deploy, but we make sure we're getting the most optimized performance out of it. I think you guys have hit something on that. I think the Hadoop thing, people were asking how come Intel's doing their own thing and not endorsing either Hortonworks or somebody else. Because of one of these types of implementation where portability is really key, people don't wanna throw away their investment, especially IT guys. They have gear, they've bought previous stuff. Maybe they can retrofit, but they don't wanna throw anything away. So having that application portability so they don't have to do rewrites is critical. So that being said, what would you say right now does this disrupt the most of, in a good way? Meaning that what Moonshot is proposing to do is pretty disruptive. What do you see the disruption areas being that are the positive disruptions for the customers? Well, I think with systems like Moonshot and they talked about that convergence of architecture which means you're gonna get a lot of efficiency out of it. It's really giving people the opportunity, I think it's more about workload-optimized systems. Really about challenging one size doesn't fit all and that we should be looking at individual applications and when you have a very large Hadoop distribution and you're running complex queries and you need a more complex processor, you may need a different system than this, but when you're not, this is gonna challenge the economics and drive that lower power, more power efficiency for those certain workloads. So I think it's really about when the system works together, saving costs and the more money we save, the more money gets invested back into new services and it's good for all of us. And interoperating with the cloud is critical. The cloud has to be there, that's obviously an economic advantage. Yeah, and absolutely and one of the things, we've met before at the Open Data Center Alliance events. One of the things the Open Data Center Alliance focuses on is that open interoperable flexible solutions where you have VM interoperability and workload portability so that you can move if somebody builds something in-house or a service provider and another customer wants to move, there should be some. Just a quick plug for the Open Data Center Alliance is a lot of references, it's a great community they're sharing. It's moving very quickly. It was kind of like a great place for peers but now with the cloud, they're sharing a lot of reference architecture. So we were there with the cube or mini cube last time. Just wanna quick touch on security. One of the things you guys emphasized in your Hadoop distro announcement with security, you're in the cloud space, obviously a hot topic. How does the security model change with this type of architectural approach or does it change? And again, what's Intel's perspective on that? Well, our perspective has been you need security at every layer. You need it within the hardware that's built within your infrastructure resources. You need it connecting your server to your client. You need it within your network. You need security to all layers. So I think this is just another step that when there are a lot of things that companies like HB and Intel continue to do that will help drive security advantages. And Moonshot is just part of their system with good management. They talk about all the good ILO capabilities and things. That's gonna help people manage their infrastructure well. Excellent. All right, Regene, thanks very much for coming on the cube, bringing the props. Good luck with the initiative and we'll see you around at other events. Yeah, thank you for having me. All right, keep it right there. This is SiliconAngle.com's theCUBE and we are here in the Big Apple New York City live at the HP Moonshot announcement. We'll be right back with our next guest, Mark Potter, who is the man today and so keep it right there.