 The HP Labs, we've talked to them in the past year about some of the Mishandra Kanhu will be on later as fellows here. They've been doing this for a while. They've been playing around with this energy efficient mindset. It's always been an HP Labs project in the labs. Yeah, so going to market is something that's been a challenge for HP Labs. What attracted HP Labs and HP as an organization to you guys? What was the one thing that you guys nailed that you can point to and talk about specifically in the technology? I think it's finally possible now. You can check the box on good enough performance for server workloads. But the aspect that really got their attention, I think, is our fabric technology and the way that you can actually scale this out with minimal overhead in terms of network costs, network latency. And it's that piece that allows you to pack these in really tight so you can be very dense. You can shrink, you showed 10 racks shrinking down to a half a rack, for instance. So that was the invention that we brought to the table that really was the final piece. So back up, let me hit the pause button on that. So you just said taking 10 racks of servers down to one. To a half of a rack. To a half of a rack. That's kind of the innovation. So for the folks who didn't see the demo because they didn't live stream it, they had this huge set of data center racks and then they powered them down to essentially half a rack. Pretty impressive demo. 2,800 plus servers in Iraq versus, say, we're talking tens of servers. I don't know, what, 30? Call it 40. Let's give them 45. Sure. OK, that's pretty dramatic. Now, each of these 2,880 servers in a rack maxes out around five watts. That's all in for the processor, the chipset, the memory, the fabric, everything. So when you do the comparisons with traditional servers, you're increasing dramatically the server count. So immediately you'll think, oh, wow, it must be way more expensive, but it's not. And it's because you're sharing resources in that chassis, right? That's right. So we've built it from the beginning to scale. So we built it with a cluster in mind, not just a low power single node. The technology is orders of magnitude lower power than what you're starting with. You need more of them to make up the gap in performance, but the workloads that we're targeting are throughput oriented and they're scale out oriented in terms of how they build out in performance. And so the net savings is significant. HP was showing something like 90% power reduction, 90% space reduction, 60% plus percent cost reduction. So, John, we have a big audience that just joined us. We have about 550 people watching. So we probably should reset and tell them that we're here at HP Labs. John, why don't you reset? OK, we're here at HP Labs in Palo Alto, California, building one of HP and behind us is the office of the founders, Bill Hewlett, Dave Packard. And we're here with the CEO of Calzada, who is the company that's on stage with HP introducing a new chipset that is just going to change the game in server performance, taking a, what's the rack size? Ten? Ten racks down to a half a rack. Ten racks to a half a rack in the server business, mostly known in the mobile business ARM, which is low energy chips, is now coming to server class scale architecture. And we're here discussing that. And you guys talk about the implications of the HP deal. What does this mean for your business? And you guys, just to be clear, you guys provide, Calzada provides the secret sauce to take that ARM chip, which is a mobile phone, cell phone chip. And then you add the secret sauce to make it server class data center ready. That's right, right? And we've been working on that for four years. OK, so this is not something that you just trained up last week. So what kind of lead do you think you have? One of the things John and I were talking about is, Intel's not here, although HP said, well, Adam's on the roadmap. What kind of lead do you think you have on Intel? Hard to say, I think based on HP's endorsement today, they must believe that there's something viable here in this technology. And it's hard to quantify, right? Because they get a big boatload of cash that they could put at this thing. But it's not just cash, right? I mean, it's not like Intel hasn't seen this coming. I mean, they've been watching this trend for a while. This is just a piece of an even bigger story. You see the ARM architecture growing up from its power-optimized orientation and the PC arena, power starting to matter everywhere, of course. So some analysts are talking about ARM-based notebooks, ARM-based desktops. We're focused on the server. So I think the architecture itself is very interesting and viable as it grows up in performance and can tackle all these workloads. And the business model is certainly interesting. The architecture is available to lots of licensees that can go often and optimize and target specific applications or specific sub-segments of different spaces. It's a hard problem you guys are tackling here. I see Intel's not even up there yet. So when you guys sat back and started designing this, you know it's a hard problem. When did you have that moment and say, damn, we could do this? And what breakthrough did you guys get through and say, okay, this is viable? Can you just share your story there? I think we always had the confidence that this could be done. We saw the future of the ARM technology and where process technology was going. Certainly when we first engaged the market, we were hearing questions back like, what's ARM, right? Nobody heard of that in the server space. So we had quite a bit of education to do and I think the market was very skeptical for a long time and then really I'd credit the Microsoft announcement in January about supporting Windows 8 on ARM and the client side as really being a tipping point that captured everybody's imagination that this will be a viable architecture in the whole PC arena. Well, what about Microsoft? They're obviously missing from the early list anyway. You've got, when you look at the list of software players here, it's the emerging Hadoop ecosystem and all the cool stuff, right? Is that Microsoft's not cool or what John and I talk about all the time? But do you expect within the next two to three years that Microsoft will jump on board or are they just playing catch up here? Well, I think the applications that we're seeing where this is a great fit, a lot of Greenfield applications, it's predominantly Linux. The application stacks are built on that. So that's where the gravity is right now for these small, very power efficient ARM-based servers scaled out. That's where the action is. We'll be at Hadoop World next week and it'll be a great event for us and there's a lot of activity going on there and that seems to be the software suppliers that are partnering with you in terms of... HP has a division, they call it the industry standard servers which is kind of an oxymoron because the standard servers have been based on old chipset mentality, right? So you look at like, I'm old enough to remember the PC revolution so remember the PC microprocessor, x86 grew up and servers came out because they were cheaper and expensive than Unix and then that became the wind-tail server growth and Blaze got smaller fast, but same architecture. Can you talk about for the folks up to the implications for the marketplace relative to this new architecture? Obviously shrinking down into a footprint saves energy, data centers are expensive, real estate and power. These are really important problems that you're solving but talk about the implications of the industry in terms of the trends that in the use cases that this is going to power. Well, I think here initially there's a set of applications that are going to take off with this that will scale linearly with nodes and Hadoop and big data type applications are great ones and Cantor Fitzgerald was talking about their problem and how they're trying to sift through today's data and yesterday's data and going all the way back they have a huge data challenge. Looking forward, when you see the evolution of the ARM technology and roadmap that companies like us can leverage, you can really see those applications and workloads that lists starting to grow and potentially over time being able to address all the interesting and relevant spaces in the server market and because you can get the power and the space way, way down, today you have something like 50 million servers installed. You can imagine 10 times that infrastructure, a billion server era if you will in the future. I think that's going to be a very interesting application in terms of what you can do with all that processing power without all the impact from cost and space and energy. A lot of people have been talking about it and Dave and I have been particularly covering the whole flash SSD market. They showed the actual product which we have pictures of and we'll get photos up on the net as soon as Mark Hopkins can get them downloaded off my camera but interesting design you have to think about the multifunctional components of servers. It's not just board, processor, you got SSDs now. What challenges do you see going forward? Obviously solid state memory is going to change the memory landscape in terms of caching and application level. What do you see as the challenges going forward to the next couple, 20 mile journeys that you guys have to take in the marketplace? The move to flash is happening now. It's been expensive and there have been technology challenges with making it work in a sort of enterprise class server environment. Those are being solved and when we look forward we see companies moving wholesale over to flash and certainly that's a huge power savings and what we see is when you take the power of the drives even out of the equation you get the power of our solutions down even further and so that's significant for us. This initial set of workloads we're very well suited for. We've got some technologies in the hopper for another generation that are gonna add to those and really extend and build out on the goodness of our fabric and then in the future Arm just announced last week 64 bit on their road map so we think that will again open up even more applications and workloads when we deploy that. You described it as an inverted pyramid, right? You're gonna have the few mega scale buyers step up initially. You guys obviously see them as leading indicators. Is that right of future data centers? They are. I think certainly there's the mega scale folks that are deeply capable all the way down to the lowest levels of software. But what also I think is interesting is you're seeing the financial services companies start to step up. Cantor Fitzgerald was on the panel today. That was surprising, right? You wouldn't expect that they would be hopping on this technology. They have huge problems in space and power and lots and lots of data to analyze and so I think Enterprise IT can probably relate a little more to a financial services company than a giant hyper scale. Final question for you is we have our next guest who are on tight schedule is for the folks out there who are in the industry who are in the weeds with the Arm technology. Arm is a hot thing. Everyone here Arm, Steve Jobs mentioned that he chose Arm over Intel for the iPad and they think of it as mobile. So people think mobile they think I want battery life and if anyone's got an iPhone or an Android phone know which one's got better battery life, right? So that's battery power to the consumer. Share with folks what, why is Arm so hot and two, what impact will it have for their lives? Just in general, just kind of a high level. So from a technology standpoint just quickly that the reason that Arm has such a better profile from a power and battery life is it was built with off as the default and you turn things on when you need it versus the other way around with the current technologies. But the Arm business model makes the CPU investment, takes that CPU investment that Arm does and it makes that available to a huge number of companies that can take that and target and optimize and build highly integrated, very efficient products for a much smaller space or segment versus a horizontal approach of a... People say purpose built, you mean like purpose built like actually, right, or products. And the goodness is that it's still architecturally consistent with all the other Arm implementations for other spaces so you get the virtue of that software development and that virtuous cycles, you're starting to see that really accelerate now. So the benefit then would be more products faster than more in place. More products, more competition, more cost competitive for OEMs that are buying them, all those things. Well, Barry, thanks for coming on. Barry Evans, CEO and one of the founders of Calzada. We have your VP of marketing, Carl's coming on a little later too, I believe. So thanks for spending some time with us. We'll be watching, very exciting and again, congratulations. Thank you. Thanks for coming on theCUBE.