 Live from Las Vegas, Nevada, it's theCUBE at HP Discover 2014, brought to you by HP. Okay, welcome back, everyone here live in Las Vegas, we're here for HP Discover 2014. This is theCUBE, our flagship program. We go out to the events and extract the signal from the noise. I'm John Furrier, the founder of SiliconANGLE. Enjoy my co-host Dave Vellante, co-founder wikibond.org, our next guest. theCUBE alumni, Jim Contier, Vice President of Global Marketing, HP servers. Welcome back again, great to see you. You look great as usual, handsome, well dressed, spiffy as they say. I work for Antonio, you can't be a slouch. He is stylish. Don't bring me down, you're looking better than I do. Antonio was on yesterday. How you doing? Actually doing really well. I mean, we're pretty excited about all of our announcements and everything that you've heard. Of course, as you can see, that line is extremely long for our new Apollo family. So pretty good so far. Let's get to the Apollo thing. That was big news on day one. Take us through some of the feedback that you just mentioned before you came on, you talked about the shareholders. There's a big buzz around this. Obviously, HP server business is not small. You guys are continuing to innovate. What's going on? What's the big news? What's the big results? Line's been around the corner. You have some guest appearances here. Including an astronaut. Tell us. Yeah, so basically what we're doing here is we're redefining HPC as we know it. We all know that HPC traditionally has been the realm of academia and government. Well, our mission is to now make that available to all enterprises of all sizes because we believe a lot of those innovations can truly help change the world to use a phrase from the deputy director of NREL. So what we announced this week was the creation of an entirely new product line. And we refer to it as the HP Apollo family. And in that family, there are two basic platforms. The first one is an industry first. It's a liquid cool system that doesn't have the risk. And here's what I mean by that. In the Apollo 8000, the way that people have done liquid cooling in the past, we all know probably mixing electricity and water is not a good thing. So what we've been able to do is come up with a way where the water is actually separated. And so you get the great performance, forex the normal performance of anything else. You're able to do it in a very small footprint because you don't have to have all of that power and cooling infrastructure. Oh, and we're able to also do it in an ultra efficient fashion. Now, don't believe anything I just told you. The best way to understand that story is to hear about one of our customers. And one of the things we're really, really proud of is the fact that we actually power the NREL, the National Renewable Energy Labs, which is a subset of the Department of Energy. We powered their peregrine system. And what they've been able to do is save about a million dollars' worth of OPEX and CAPEX every year. Oh, and get to work on really cool, fun things such as biomedical research, climate control. And the next time that you see a rover land on Mars, those solar cells were designed on that platform over there. And this was a technology, we had John Grimala on the LN and they were saying that it was a collaboration between the R&D group within the server division and HP Labs, is that right? It was a collaboration of HP Labs. It was a collaboration of our team. We also had great collaboration with the folks at NREL. We built basically over 18 to 24 month period, two prototypes and we both learned a lot on it. As a matter of fact, a lot of those refinements came into place and that's where we really shine. I mean, working hand in hand with our customers to ensure that they not only get the best possible performance, they're able to get the most energy efficiency and they can do it in the smallest footprint. One of the other things that was really funny and interesting is as we all know, high performance computing systems throw off a lot of hot water. So as we were kind of having these conversations with the folks at NREL, because their mission was to build one of the world's most energy efficient data centers, they were telling us that they're using that hot water to eat the sidewalks during the winter. And that's again, OPEX and CAPEX that they don't have to go spend any money for. Save on snow plowing. There you go. So how much of the, you mentioned that HPC has traditionally been this niche, an important one, but I don't know, 10% of the market, maybe even single digits, how much of the enterprise market do you think this type of technology will be able to capture over time? Just how much is even a candidate for? So the good news is if you're looking at, instead of just looking at it from a market share perspective, what can people do with it? So if you're doing computer aided design, the Apollo family is a really good way to do some of those kind of things. If you're looking at everything from genetics to medical research, I mean, we had one customer, why can't I? Who wanted to calculate the amount of air and the turbulence of the air around a product that we're going to basically, let's just call it launch. So the net is, is that to answer your point, today, those kinds and classes of customers are what we'll call like 10 to 12% of the market, but with advancements like Apollo, we want to make that available to enterprises of all sizes. So probably within three or four years, you might actually see that almost 15, 20% add in terms of the overall market. Aren't big data workloads a candidate for this type of computing? Is it sort of big data meets HPC or? Potentially, but where this really shines, and I'll answer your point. So big data, we've got a great little product for that, that's referred to as the SL4500, which you folks have talked about. In essence, what that product does, it's a true, it was the first server design basically for big data. And what we did is we looked at just the perfect amount of cores, spindles and IO. So to give you an example on that one platform, we can hold up to 240 terabytes of data in basically a 4.3U area. But to answer your point specifically, where this product really shines, is when you're actually doing genetic research, when you're doing climate control, when you're actually doing finite elemental analysis, simulations, modeling, car design, that's where something like the Apollo family basically really comes up into play. Now, I've been talking about the 8,000 for most of this, but there's also, I'll call it the little brother of that, which is an air-cooled system. And the air-cooled system is the HP Apollo 6,000. The lunar module. There you go. Yeah, it basically detaches and comes down. No, but kidding aside, back to your point of where do these things play best. That's a perfect example that if you're looking at engineering design automation, if you're looking at modeling and simulation, let's just say Monte Carlo, Black Shoulds, or a large financial services company, Actuarial, where you need a lot of compute power and you can do it in an air-cooled fashion, there's 160 servers in that one rack. And in that 160 servers, to give you a really good customer business example, Intel, Intel was our very first customer for this. So Intel Silicon Design Engineers are now using a combination of our product and their product to design the next class of Intel processors and architecture. That's a circular reference right there, that's good. Now, you had an astronaut come in for the launch, you know, pun intended, talk about that from a marketing standpoint, what were you thinking there? I mean, obviously there's a connection there, but what was that like? Actually, it was really cool. So we had, and I'm going to use the words, the honor of hosting Dr. Story Musgrave for those that don't know him. He basically is one of the most, he's one of the most, I guess we'll call it, most missions astronauts out there. And as a result of that, what we were able to do was not only host him, and he came and talked from the heart about innovation and what happened back in his days as an astronaut. He talked about what it meant to be a risk taker. He talked about what it meant to think different and to help change the way things were being done. Now, he's a pretty accomplished individual, right? Give us a little, tell us about it. Talk to Musgrave's background. Yeah, I mean, basically the things that he's pulled off in his lifetime, there are multiple people that don't pull him off. So to give you some succinct examples, the gentleman holds seven degrees. He's got 20 doctorates. He's actually a true medical doctor. He has had 800 jumps out of perfectly good aircraft as a parachutist. He's actually a commercial pilot. Oh, and by the way, for fun, he basically is a post trauma surgeon when he gets in free time. So I mean, this is one extremely, not only impressive individual, but somebody who's also very down to earth and he's funny as heck. Humbling, hanging out with those guys. So the line around the corner, so David, I've been watching your booth right here. So what have been some of the customer feedback? Because like, are people blown away? I mean, what's the general reaction? Because it's been pretty deep here. You had the big reveal. Oh yeah, no, it was a huge reveal. So on Monday, we had the full press play. On Tuesday, we actually had an on main stage with Meg and Bill. And then as you can tell, we had the reveal here on the show floor. To answer your point, people are blown away because most folks understand that when you typically do high performance computing, it costs a lot. It takes a lot of space and there's a lot of power. I mean, the guys and I were doing a couple of quick calculations as to the linear trajectory of high performance computing. And we said, let's look at the top five length list. If we were to extrapolate that list over the last 10 years, it's grown roughly 2x. But if we were to extrapolate for the next five years, you would need basically a million terraflops in order to pull that off. The energy alone required, you'd need about a gigawatt of power. Let me put that in perspective. A gigawatt of power is the output of the Hoover Dam. Or as the puppets would say, John, a gigawatt. Yeah. And then worse, you would need 30, and I'm from Texas, right? You would need 30 football fields in order to house that piece. So when they come and they see how much performance we've put in a rack, the innovation around the power and cooling for the water wall and what we refer to basically as our thermal bus bars, they're just basically blown away. As a matter of fact, there have been a lot of customers who said, hey, how do I get a couple of these to start testing and validating and how do I actually get kick-started on this? So they're salivating basically over it. Well, I can tell, you can see the line right now. It's still kind of wrapped around the corner. So let me just kind of break this down. What are they most blown away by? Is it just the densely, is it the power and cooling, footprint, all the above, they break it down. Yeah, basically what they're blown away with. Yeah, so what they're blown away with is the fact of with this increase in performance, it doesn't require all the power and cooling that you would normally meet. Because as you can imagine, when these things are running, these are high-performance computing platforms, right? They're throwing off a lot of heat. Well, traditionally you gotta go spend a lot of money to cool that stuff down. Oh, and by the way, you've gotta spend a lot of money on all of those motors and fans and crack units and everything else. So they not only can get blown away by that, the fact that we can fit 144 computers in that one rack because we took a rack scale approach to it, also blows them away. And then last but not least, it's the fact that we can put it in a small footprint. Now, I keep talking about performance, right? Again, just to put it in perspective, that one rack that you're seeing over there, which we're delivering today, if you had taken the top 500 supercomputer list last year, we have at 250 teraflops, the equivalent of what would have been the 158th or 159th fastest supercomputer in the world in that one rack. So really good performance in size. So Dave and I were talking about this a couple of weeks ago, putting together our view around HPC, our research agenda and our editorial agenda and really looking at the changing HPC landscape with respect to clouds. I want to get your take on it. The on-prem cloud integration plays well. Does this fit into that? And what is the vision for high-performance computing? Because Antonio said your innovation strategy has been clear as day from day one. It's compute. The right compute for the right workload, the right economics every time. It's a compute vision. Correct. Compute is the key. Glades, all that stuff will still hang around. But again, the day clouds right around the corner. How does that intersect? What does HPC cloud mean to you? So it means a couple of things and I'll actually break it down into two distinct components. The first one is I've been telling you all about the product. In addition to one of the other announcements we made here was the capability or the availability of our HP Helion service for HPC. So again, making it available to enterprises of all sizes. If you want to actually have this as a private cloud on-premise, if you want to burst in and run HPC level applications, our Helion offering can actually do that for you. Second piece is that we're finding and we have the privilege of working with a lot of universities. A lot of them are now looking at this and saying, hey, you know what? Instead of each of us standing up at a high performance compute center, what if we created a regional center and then we could allow others to come in and utilize it, you know, basically using a cloud-like optic? Those are the kind of things that we're seeing both in terms of not only the very fast growth of HPC. Matter of fact, if you believe some of the prognostications and you and I both know that not all market prognostications are correct, some folks are saying that within the next four to five years, HPC may actually be anywhere up to an including about 40 to 50% of the market. So, fast growth, lots of requirements for performance, very little requirements for power and cooling, big requirements for small footprint, your prem, our prem, or a hybrid or converged scenario, that's what we're hearing from customers. Great answer, because it's our number one question that we get and we're trying to solve that answer. Solve to get that answer. So I got to ask you the next question. Who is the target audience for that product? If I had to like create a data model around looking and understanding who would be the right person to buy it and be interested in it, who is the persona, who is it? CIO, Facilities Guide, Data Center Guide, your server, buyer, I mean, who's your target audience? So we actually have a couple of personas, but the primary one would be either the vice president of IT or the vice president of operations because these are people that have to meet service level agreements. These are people that have to actually hit those performance price and operational expenses. So those are the two that we would primarily talk to. The other group is, believe it or not, we're getting now a lot of line of business owners who are partnering with the CIO, who are saying, look, it's not really about the technical requirements, it's about enabling the business. And that's one of the fundamental precepts we've put in place with this making it available to enterprises of all sizes because we want to advance innovations from what typically would have taken years, you know, make them as fast as months. So the DevOps culture, you're seeing that kind of notion of application-driven workload-driven purchases, meaning I got an app, I need compute. So it's just like, you guys have been suppliers in a lot of these big hyperscale companies you probably can't say the names out because of the confidentiality, but they're buying like massive, it's no like speech-in-feeds, it's like, give me, deliver all this product. Yeah, and by the way, there's some great logistical requirements there. I mean, you guys were, have been with us since the very beginning and you know the pod store very well. Case in point, if you want to deliver a huge amount of compute storage and networking or basically a two-data center, not only do you have to deliver it, but sometimes we get requirements to say it must be delivered on this date between these hours because that's when we can actually do the install and take the systems down. You're the engine of DevOps, basically. Pretty much. In my mind, if the DevOps trends continues to trajectory forward, that's an application developer mindset that's going to have virtualization requirements, workload requirements, app specific. Yeah, and by the way. That's the business line. You're dead on and the Apollo 6000 is a perfect example of that. I mean, the application is silicon design. It's going to turn it into a real-world product and if you need a lot of compute with very little power draw and a very small footprint, that's the box you want it to basically, or I should say that's the rack that you want to go take a look at. So what else is exciting? You're right. We used to talk to you about just ISS. Now you've got the scope of the entire server. Good business. You've got Moonshot. You've got the hyperscale stuff going on. You've got the machine, the matrix. No, we don't have the machine just yet. Yeah, that's still in the lab. Yeah, but it's good. It's good. There's a great collaboration. It's good for being relevant in the future, right? So there's great collaboration. Tell me what you're excited about. So actually, we're excited on a couple of things. We've been seeing really good progress in terms of the acceptance of this new compute strategy. We're really excited about the number and class of folks that are starting to basically adopt Moonshot. As a matter of fact, we've got a combination of customers such as IncaVenca, MyLock, and various others. eBay came out the other day. That's correct. Yeah, so as a matter of fact, I was just with Dean Nelson over in another studio and he was basically doing a testimonial on our behalf. So we're seeing a lot of really good progress. And again, it's not because it's a great platform. It's because we really are going after our customers' biggest needs. We're helping them reduce their power. We're helping them reduce the footprint. We're giving them the kinds of density that only Moonshot can do. So that's working out pretty well. HP Blades, plus one of you, great take-up. As a matter of fact, I'm sure you've probably heard this throughout the entire week. Software-defined data center, software-defined networking. If you don't have something like HP OneView, you really can't do it because folks are moving away from individual boxes of server storage networking and then even underlying services to now an integrated model where when they want to deploy, according to their service level agreement, they're deploying them rapidly. And when they want to do updates, manage, move, or maintain, they're utilizing it all with one common management construct. What about the Foxconn deal? What do you tell us about that? Actually, so the Foxconn deal, which was announced on April 30th, is our way of addressing the very fast-growing service provider space. And what's happening there is back to what you were mentioning, we've got a partnership with Foxcon with what most folks may not know is the world's largest electromechanical supplier. They do work for us. They do work for Apple and various others. It's a great complementary partnership because we're going to be able to take a combination of our engineering, both of our supply chains, their design capabilities, and as you can imagine, from an economics perspective, be able to deliver at massive scale to service provider-class customers that are interested in just basically providing that level of compute. And what about the machine? What do customers talk to you? When they see the machine, they say, ah, that's nice. No, actually, yeah. They say, hey, I really like the direction that you're going. Depending on the customer. So the customers that are, you know, a little bit more advanced, they look at that and they go, wow, I get it. This is basically the direction the industry is going to move to. How fast can I start looking at implementing portions of this, whether it's the memrister piece, whether it's the photonics piece, whether it's a silicon or I should say, SOC piece. So we're getting a lot of interest. Now, as you all know, that's, you know, a couple of years away. But the good news is, is that it's resonated. It's hit a chord. And now, for some of our more advanced customers, they're already starting to hit us up and ask for what would be the equivalent of a beta. Jim, I got to ask you, because last week I did an interview with Cisco, and they're claiming to be number one in servers in North America. So they came on the records on Cube interview from last week. Todd Brannon and Raghu Nubiar, engineer over there. And what's your take on that? I mean, you're a market share, you're all over it. And Yossi Antonio Neary, he's pointing out the global numbers. Give us the update on what's going on with the IDC numbers. They seem to be all over the place. So I don't know who to trust at this point. So tell me, I trust you. Tell me what you think. So let's start at the very beginning. That claim. I'm sure you guys know the old Jack Welch phrase of you must be number one or number two. And one of the outputs of that is people were sub-sectoring markets so small to the fact that they were the only one left. Now, it's not as bad there. That claim is not 100% accurate. I tend to look at things at a global level. And if you look at Cisco share, the Cisco share at a global level is in the single digits. And when I say single digits, you can pick your favorite number, either 3% or 6%. Having said that, again, that claim, it's a single market. It's a single platform and it's a single metric. It's revenue only. And one of the reasons why it's revenue only is somebody should really go back and take a look at the way that they're calculating some of those average selling prices. I don't know about you gentlemen, but if I were gonna buy something that people would see as perhaps the technical equivalent, being charged 40% more for it, probably wouldn't make me feel really good. So I would have those gentlemen go back and not only have a conversation with their prep teams, but make sure that the data itself is really reflective of what's happening at a global level. When I talked to guys in the channel, and granted I talked to, recently I just did a little channel survey and I was talking mostly to the storage resource. They told me that essentially, if the price is not within 5% to 8% of each competitor, they won't sell it. So when you see a 40% delta, you should question the number. So I was gonna ask you, is it sort of similar in the channel and servers or is it even tighter maybe? Depending on the product, depending on the platform, it does vary, but you're within the right ballpark. If you've got some sustainable advantage maybe, but generally across the board. Correct, so I don't understand, Dave, you worked there. You basically built that whole group out of scratch and you work, you know the numbers. I don't understand though how they can claim that the number one, I mean, Michael Dell said Dave, he was number one. And that was not accurate. So I can say the number one in Harvard, Massachusetts. Correct, sub-segmentation. Number one. So they're bogus numbers from your perspective. The, no. They're numbers. They're numbers. The problem though is that again, when you're looking at a single market, single segment, single platform, one metric, then yes, congratulations. Well I think what's happening is it sounds like the methodology is including revenue from all the piece parts that they're calling it conversely. What happens a lot in this market, I've noticed this, somebody will install, let's say, let me just use a NetApp example. Somebody will install a NetApp box, or let's say, no, sorry. Somebody will install a UCS and maybe three or four months later, a NetApp box will go and they'll call that a flex pod. Correct. Okay, that's, they'll call that a converged infrastructure. Is it, I guess, it could be, I mean it's a reference architecture. All that revenue sounds like it's going into the numbers. And that's why there's a 40% premium. I would tend not to like to count things that way. I'd like to see it more segmented and still have the piece parts or put it all into a converged infrastructure bucket. So I think that's probably, take some time for these things to work out and we're a big company with a lot of. But to answer your point, we need to shine the flashlight of truth. Single segment, single market, single platform, one metric out of the multiple types of metrics you can provide. So you guys are number one. Worldwide, in EMEA, in APJ, on multiple items, including mostly units and then almost all of the revenue. So for us, it's not just looking at it with a singular optic. It truly is a global business with global results. And to me, revenue's the most important. Assuming it's an apples to apples calculation. Correct. You and I both know that we're not sure it's 100% apples. No, we're sure it's not. Actually, apples to tangerines. Okay. Chip Gonti, great to have you inside the Cube again. Cube alone with one of our favorite guests. Thanks for coming on again. Clarifying and shining the truth, the light of truth on the IDC. Truthful numbers, but like kind of just categories differently. The Jack Welch model. The Jack Welch model. And you can be number one in anything. Correct. We're number one in Cube broadcast at events. Congratulations. In the sales convention, sir. There you go. It's you. Great to have you. We'll be right back after this short break. Thanks guys.