 H.P. Discover in Barcelona, Spain. This is SiliconANGLE and Wikibon's theCUBE, our flagship program where we go out to the advanced extract to signal from the noise. I'm John Furrier, the founder of SiliconANGLE. I'm joined by co-host Dave Vellante, co-founder of wikibon.org. And we are here with John Grimala, who's the senior director of hyperscale product. Manager. Management, management. Which is essentially the large scale, stuff that we have been talking about for years, Dave. The large scale data centers. We're talking about massively scalable systems, things that Facebook and these guys would want to buy, large enterprises, which are now moving into that world with open source. John, welcome back. You've been a CUBE alumni many times. We've covered the Moonshot launch. Moonshot pods, this is the future. Tell us, what is the update and how much is hyperscale truly now in the enterprise? It used to be kind of like an anomaly or a black swan in the sense that Facebook, Google, and the Amazons are doing such huge scale. Now the enterprises are now moving to that. In that direction. To what level are you seeing that? It gives us an update. We're actually getting lots of interest across the board in this. New products like our SL2500, Proline SL2500, make it easier for people to deploy at this kind of scale. This is a simple 2U high box that someone can get a four node system and go ahead and deploy it. We've also got tremendous interest in our SL4500, which is a Proline system that can have a maximum storage of 60 drives internal to the system. So think of a rack of over two petabytes of storage and you can put whatever storage optimized application you want on that. Great opportunity for object store, which is a great piece. You can also go with typical type of exchange or Hadoop as well. So lots of options there. Everything's looking cool. So we last talked about with you. I was just pulling up the video. That was a little sound from our video in New York. We were in New York City. We had the whole team. Really it was the Moonshot launch. Again, we've covered the original Moonshot two years ago at Bill and Dave's office in Palo Alto. Cube, Cube Technology Day there. Give us the update. Moonshot is obviously the hottest. I don't want to say window dressing because it looks like window dressing when Meg kind of uses it on stage. It's such a high lightable notable thing, but it's not yet in the major mainstream of the HP dollars and revenue, but it is hugely important. And Antonio was drilling down on that. Give us the update on Moonshot. What's going on? What's the ecosystem look like? What's the update? What are some of the customers? Can you share some data? It's a really revolution. It's actually been a very exciting year. We started again with April, our initial launch of the Moonshot system. Moved on to actually do a lot of partner-centric announcements. We spent September going through announcements with Intel and their Aviton parts, with the ARM TechCon and each of the partners that we announced with AMD at their partner conference, and again with Texas Instruments. So we've been on the road working with our partners closely, which is what we talked about in Moonshot, that vision of how we innovate at a faster pace overall. And we're real happy here at Discover to go ahead and announce that we're shipping the Proline M300. This is the Aviton-based server cartridge. Really a more compute power capability. Really one of those servers that everyone's been waiting for. This is, again, Aviton-based. This is a processor that has roughly the same performance as servers that we had about four to five years ago. So think of our generation six servers that you can get that on a Moonshot cartridge now. In addition, probably the biggest interest we've had is in the hosted desktop area. This is based on our Proline M700 AMD-based server cartridge. Huge, huge interest in being able to do HDI hosted desktop solutions. We've got a great package solution around that called the Converged System 100 Hosted Desktop Solution. We've had people three, four, five deep wanting to see those demos. So a lot of people are going to say, well, wait a minute, John, you're taking me back to gen six, why would I want to do that? So explain the concept. I mean, we're talking to cartridge, we're talking way less energy, we're talking scale out. Yes, exactly. You're getting, thanks for that. So exactly you're getting at scale that level of computing power. So you go back to get 89% less energy, but all of the performance you had four years ago. And I can put 45 of those inside a single for you system with Moonshot. So the density of computing I'm getting is at a scale that I haven't had before. So I can dial up my compute capability. Absolutely. So if you put it into perspective, that box that we've got for Moonshot, if I put an aggregate what I've got there, I can get 45 terabytes of storage, plus 360 cores, plus over 1.4 terabytes of memory across those 45 cartridges. It's that I've got them spread across the cartridges versus one single server now. So tremendous capability, great alignment to a lot of the scale out applications. We talked to Tom Joyce yesterday, and we heard from Antonio Neary today, and both of them were focused on emphasizing segmentation. And we were asking, when you talk to Meg, what he talked about is it really drills us on segmentation. Of course you came out of a consumer background, so segmentation's very important. So I want to talk about the market segmentation for Moonshot. Now, you're creating a market. Absolutely. But, and you're disrupting markets. So how do you look at the market segmentation for Moonshot? Yeah, anytime you hit a new category like that, you're blazing new trails, so to speak. But it's an extension, if you want to call it, of our density optimized set of products or our hyperscale products. We're building and establishing sort of a new frontier for that. But we're expanding that frontier out of what we're just service provider type folks, and HPC like customers, well into the enterprise now. When you've got something like a hosted desk type solution with our Converged System 100, it's something that's extremely easy for even an enterprise to go ahead and deploy at scale. And this is a great example where I can take basic things like websites and a lot of hosting and hosted desktop kind of things and make it available at scale for enterprises in a very easy to deploy. So I was, I think I got an email from Stu the other day and it had some registrar, I think it was a registrar article on IDC came up with some new numbers around hyperscale and ODMs and saying, you know, talking about how much share ODMs are taking the market. I think it was like 10 or 15% or something like that. So it's starting to get meaningful. Should we think about Moonshot as going after that market segment or is it not the case? It's really sort of- Well, that's where the- High density for the enterprise. Yeah, that's where the whole concept of a software defined server really matters that I can sort of change the personality of that server depending on which of those market segments I want to go after. If I want to target telecommunications areas, I can configure that system with our Texas instruments, cartridges or some that have special offload engines to go ahead and deal with media transcoding and all. And it becomes a personality of a telco focused box or just general hyperscale or even enterprise with things like hosted desktop. So what's awesome is I've got the architecture that can support the scale, but I can go ahead and configure it for the solution that I need. One of the things that we've been talking about obviously the success of three par and storage has been the performance of the storage group and just the things that they've been doing is amazing. You got now Tom Joyce, swimming with the sharks with his new program, the Converged Infrastructure of the Shark Program, Shark Tank, we call it theCUBE or Shark Week here at theCUBE. All these trends point to the integration. And I want to talk about this data center concept we were just talking about earlier. What do customers need to understand when they want to start moving to this fully converged basically ordering a data center or we've talked about data center operating systems in the past, fully integrated software, hardware, power and cooling, really app driven. It's a new dynamic. Customers are ordering data centers and they're looking at it from an app perspective and it's all the ingredients that was traditionally piecemeal, networking, the three par storage and HP storage and servers now cobbled together with power and cooling. Talk about the dynamic and share with the folks why this is important and what are some of the examples of people and why they're buying the data centers. Yeah, what's exciting is people are taking what were sort of tip of the spear kind of examples that took place with a lot of the larger hyper scale customers and they're now looking to do that at even enterprises and a lot of very mainstream type of companies. They want to have this controlled environment or control system that you can get with our performance optimized data centers and make it easier for them to go ahead and deploy full data centers at a 13 times faster pace. So going from multiple years to deploy a data center to just a few months. And the great thing about this control system is it can be done for half the cost and half the power. Because it's fully integrated, because the cooling's integrated, not only do I get it faster, it's more efficient and people are clearly seeing that trend and the benefit. What kind of number, I mean, I'm sure you're not going to share the numbers, but give a taste of the stage of development of this trend. Obviously, we all know the large hyper scale guys are buying it, but the enterprise would traditionally didn't have the expertise to do all the integration. You guys now fill that void. What part of that market are you feeling? Is it more the mid-range, hyper scale? What are some of the profiles of customers that you are looking at? And that's actually one of the most exciting pieces with coming here to Europe. We've got a European targeted set of performance optimized data center offerings. This is, yeah, it is specifically done for the market. There are things like environmental doors that you can put in so you can not essentially have any of the outside elements that you're bringing into the data center. There's additional doors inside. There's additional ability to move the IT equipment back and forth. All of those pieces coming together make it more easy for anyone to go ahead and consume these things and deploy them. And what I've noticed, and I don't know if you've gone into some of the different airports here in Europe, they're using pods for all sorts of different things. At a general scale, the Frankfurt airport's a great example where they're using this modularity even in how they're expanding and managing their infrastructure there. And IT slots in perfectly into that with these pods. So the interest has actually been tremendous here. We're seeing more people actually wanting to jump in and buy these things than we've seen in a long time. And we really hit the sweet spot with this market. We were in 2010, John, we saw the unveiling of the pods. Remember that the press conference that HP had that you and I attended. So, and that was early days. Are we now just seeing the go to market of that concept? Or has there been changes since 2010? Yes, so the biggest change that you're dealing with is those early ones were more an American-centric first, again, with service providers. But moving out into a lot of the industrial sort of military applications and any place where people needed to be highly mobile where they did that. But now things are taking off to where we've got pods that are regularly deployed on rooftops of buildings of companies. And the roll-outs are really changing to a different scale. And whether it's a 20-foot pod, which is the one that we have here at the show, or one of the 40-foot pods which tend to be done by a lot of those service providers who want large scale, just tremendous example where that's rolling out like crazy. John, I want to talk about some questions we got. So we did theCUBE on Tuesday night here, and then we had a virtual event at the Gartner Data Center Conference, which is going on this week in the U.S. as you're aware of. Hashtag Gartner DC. And we had a flash mob, a crowd chat, and we had some commentary. The pod kind of concept came up and we were kind of riffing on the AIDS Center operating system, which I always promote that because I love that concept. But the question came up from someone that said, Dave Riley reveals Bank of America's three major tech goals. Adopting open software-defined infrastructure two, streamlining the data centers driven by industry benchmarks, and three, driving innovations. Again, high-level goals. What's your comment on that? Obviously open, it's been something that HP, are you guys truly open? You got virtualization in here. You got, you know, at the hardware layer of their issues. What does that all mean? Can you talk about those three things? I think this is one of those where we can go check, check, check. We've got all of those covered. Let's take your first one about open software. A lot of our great cloud OS work, and we're demonstrating that here again at Discover. Tremendous thing that we were able to build on top of OpenStack to get a full infrastructure as a service running on all sorts of different types of systems, including Moonshot as well. We've demonstrated that. So we've got that performance optimized data centers just like we talked about. Being able to buy and deploy at scale, automate that to get greater efficiencies, our huge opportunities. And in terms of pushing that innovation, that is the core of what the vision of Moonshot was all about. It's how we work closer with partners, whether it's AMD, Intel, Texas Instruments, that great Armat ecosystem that's out there. How do we pull partners in closer and innovate faster, bring customers into the Discovery Lab that we've established, all definitely rolling and flying? Okay, so some other commentary, obviously we were talking about IT transformation, and that led to Jeff Hewitt at Gardner had presented this talk. Everyone's coming to this new model. But then Open Compute came in. So Open Compute was a trend, which is familiar with the Open Compute initiative. It's just buy your own and kind of cobble together. So the question is, will the pods eventually have an Open Compute component? And then some commentary here was, infrastructure is always the last to undergo real transformation, which is kind of true, but does price and Open Compute Trump processing? These are the kinds of conversations that was being said at Gardner that we were having on our crowd chat. What's your comments on that? Because Open Compute is still early, but how do you guys look at that with a pause? Because it seems like a natural extension to just throw Open Compute in there. It's however you can get that optimization to fit a customer's needs. They ultimately want to go ahead and get what fits their applications and runs it fastest. And sometimes picking something that's basic and vanilla to fit that will work. There's other times when I want to put something that's an accelerated piece in there that's going to run faster. And more often than not, when we start down the road with a customer, they're really wanting that optimization to get the extra performance that's needed. When they see some of the things that you can do with different offload engines or different accelerators with moonshot type things, that gets their interest and toggles them over. So we'll support whatever they want, ultimately, but they have that choice of pulling something that's a little bit more accelerated and innovated or something that's more basic. How does someone order a pod? I mean, I'm fascinated by this because I just see this as a future in buying these data centers. So do I just dial up and say, okay, we got a simple part number at NICE. I mean, no, but it's simply like, it is a pre-canned, well, an order, but it's simply just filling out like, well, you got to define your servers. I mean, I'm the application workload manager. I know what I need. Absolutely. How do I go to the HP store? Who puts it together? Does HP put it together for me? Yeah, there's two options. First, if someone wants to buy an empty pod because our pods are really ahead of a lot of the competitors and their capabilities, we'll absolutely ship that out and people can integrate whatever they want into it. Very easy to go ahead and deploy things. It's just like racking and stacking regular systems. But beyond that, and my office actually overlooks what we call pod works, which is an area where we can take those pods and take the racks that we put into our factory express where you order the racks fully configured to your specification. Whatever you want put in, those racks get rolled directly into the pods and the pod is tested at scale and we'll ship that directly out to you. You can choose whichever of those paths that you want and we're happy to go ahead and ship that out to us. Just to clarify, John, you were saying that you can buy them empty and you said because our competitors are not, you meant your pod competitors or your server competitors? Our pod competitors. In the world of pods, the pods have really evolved over the years. They started just being sort of almost like shipping containers where you put a few things in. But if you look at the pods today, they've really evolved to where the walls are optimized for sound insulation, the power and cooling systems are optimized. Heck, we even have full cooling systems and power generation options now that you can put and connect up to those systems. So they're fully advanced data centers now that can get deployed at scale. So Dave, at the Gardner Conference, we were talking about this, I remember, I want to bring this up. Don Duet from Goldman Sachs mentioned modularization as a new data center strategy, which we had a discussion on. And there was some commentary. Not everyone's buying pre-configured buildings per se, but this talks to the pod model. I mean, this is kind of what these Goldman Sachs, obviously a big-scale company, but others like Goldman Sachs, Bank of America and others are huge companies. Well, in that scale, companies like dealing with, there's a lot of folks that we've dealt with who want to, what they called a, they put a whole row of computing in and optimize it and change it over time. Now people just want to buy a full pod, put it in and deploy it, and then three years later, when they want to head and rotate that, they'll roll in another pod and take that one out for updates. So you hear a lot about the Software-Defined Data Center. What about the Software-Defined Data Center? That's what it is, absolutely. Is that here today? I mean, can you control the flow of air and power distribution and the like in software? Is that coming? I'll show you a great, and yesterday I was in the pod with a few of the folks and one of the guys on the floor was playing a joke on the guy who was in there. The doors were closed and he walked over to the power system and turned up the fans to 100% and their hair was flying back and it's all software controlled completely and you put it on either auto where it'll run or you can move it up and down as you want. It's a full integrated control system with a flat panel that you could monitor that whole thing with. That's really when you come up with these highly efficient control systems, that's when you're getting that savings at a different level. We're talking about 50% lower energy costs, 50% lower deployment costs. Those end up being a huge thing. Let's talk about the market opportunity. Okay, so one, let's assume we just talk and you believe that what we said is true, modularization is going to happen and it makes sense with Moonshot and footprint challenges, density and power and cooling, no brain or gardeners talking about it, this is happening. So congratulations, you're vectoring down the right path. There's a market there. So how are you guys stacking up against the competition? Obviously, we Dell has some containers I know or pods if you will on their own, a little bit different. There's a company, IO data systems has mixed pods and I don't know if anyone else is in this business. We're not used to do it. What is the competition? What does the market landscape look like in the competition? And how do you guys compete? What's your competitive strategy? Yeah, there's a lot of different segments in here than this is back to the segmentation point earlier. We're needing to go ahead and target, there are the traditional service providers who want sort of basic capabilities at large scale. There's enterprises who want to start deploying these things. The Goldman Sachs, as you had mentioned, there are military areas that have their specializations. There are high performance computing areas. Each of these segments has their own special things that they want, the degree of cooling, how much they want to push the limits on the cooling and even if you need any sort of special things for dealing with areas that you're going to drop pods out of airplanes or any other areas, each of those has unique needs that we can go ahead and support. So it's all about segmentation. So are you guys going to differentiate on what? Just industry standards being open. I mean, a lot of hosting guys are getting into this because I mentioned IO data systems. I mean, IBM bought soft layer, is that a potential threat to you guys? I mean, the hosting market, this seems a natural extension for hosters. So it can help the hosters run more efficiently in the way they deploy systems. Again, it's what scale that they're dealing with. What's great, the pod that we've got here, the six meter pod is 368U that you can deploy for whatever you'd like inside that data center. And it's great that it can be compute heavy or not, and you manage it accordingly. I think it's going to make it easier for people to go ahead and create these types of hosting environments. They can be hidden on top of the roofs of a building or just simply put inside a building. Well, it's early days. We'll see how you guys compete. We'll be watching, certainly. We're big fans of the pod. So how should we think about this market? I mean, how big is the data center build out market? I mean, it's got to be enormous. It's a real estate. You guys are all real. It is absolutely enormous. A real estate guys don't know tech, Dave. That's communication. So this is my question I want to ask John, is what's going to happen to all these assets that are assumed to be, I don't know, 30 year life cycles, and the market's coming in with these modern pods? What happens to all those data centers? Is it OPEX? No, no, it's a matter of how, remember, we're dealing with this just massive pace of growth. You know, whether it's, the pervasiveness of everything that we're dealing with with cloud and internet has got a growth path for IT that people aren't sure how they're going to deal with it, but what they can do is they can use these efficiencies to deal with the growth without having to build new data centers. They'll modularize those data centers. I guess what Dave's asking is there a financial implication of the write-off factor of all these assets, end of the life thing. And the big company has, assuming a 30 year life for its data center, and they're into year 10, and then all of a sudden this pod thing starts taking off and it's supporting hyperscale and they're looking at their data center going, wow, there's no way this thing has a 30 year life. They bought it, they own it, they ride that pony, until then, you know, until then. No, I think it's a matter of how the concepts get integrated, that these are sort of pods that you'll put on a truck and you can put any place. But I think the modular data center is coming to traditional data centers, where people can actually take pod-like portions and house them in traditional data centers, sort of contained areas within data centers. That's the kind of commentary we have on theCUBE, thinking outside the pod. Yeah, outside the pod, I like that. Yeah, outside the box. That sounds like Franken data center to me. So, okay, so let's move off the pod. I want to, when we've got Jim Gunthe coming up next, I want to get some other points in. We're going to be watching that market early on. I think it's going to be exciting, it's going to be fun to see who enters that market and who dominates. You guys certainly have the muscle to do that and the customers. Let's talk about back to Moonshot. What's happening, what's next? Share with us some, any good customer announcements to share? I noticed there wasn't a lot of discussion with customers here on the keynote. Meg didn't really talk about any customers of Moonshot. Can you share some notable customers? Oh yeah, actually, great. I actually co-presented my Moonshot session with a customer. We've got, server condo was out here. John Brown, their CEO, came out. He was a very early adopter and partner with us with Moonshot where this is where we work with those Silicon partners and software partners, pull in those Lighthouse customers. And John's a great example of someone just with a tremendous hosting background and wanting to roll things out. He was going down the path of, hey, I needed to optimize. He was considering designing his own systems as well. Once he saw Moonshot, it was such a great fit for his needs. He jumped in and used that to accelerate his business. And his simple math, when he looked at it and said, I can get $100 a month for every one of those systems I deploy, I can get a rack full of half-width servers and I can get 80 of those per rack, or I can get a rack full of Moonshots and put 450 of those in. The math is simple. His revenue in was much higher and it really helped accelerate his business. So, great example of that. John, great to have you back on theCUBE. CUBE alumni, Moonshot. We've been following Moonshot from the beginning. We're looking forward to bringing the CUBE to the next event and an exciting unveiling of new stuff. It's just innovation's phenomenal. We fell in love with the product from day one. We actually were clamoring for a box. Just the overall footprint, we're huge fans. It's great engineering. It's like the old HP, like Meg talked about on her keynote. So, great to have you on pods. Very relevant, solves a real big problem. Relevant problem. So, I want my final word for this segment for you. I wanted to give you the last word is put the bumper sticker on the car leaving Barcelona, okay? What's it say for you and your team and the overall vibe here? What's on the bumper sticker? You know, I think ultimately it's innovation is going to drive us to meeting the needs of the future. And when customers see the path of innovation, whether it's Moonshot, whether it's pods, whether it's new storage optimized systems, each of those really charts a new path with a new value that we can provide for customers. And they look at HP differently in terms of just aligning with how we fit their needs. So, innovation's a great thing that helps steer the ship in the right direction. So, just to let you know, I just posted on Facebook our photo from New York with the team from Moonshot, smiling at the moon, you guys are howling at the moon, good stuff, it certainly is a Moonshot. Great to have you on theCUBE. We'll be right back with our next guest after the short break. This is SiliconANGLE and Wikibon's exclusive HP Discover coverage here live in Barcelona, Spain. We'll be right back with our next guest.