 Everybody, this is SiliconANGLE and Wikibon's the Cube, our flagship program. We go out to the events, extract the silly from the noise. We're live here in Barcelona, Spain with exclusive coverage of HP Discover 2013. And we're back fourth year now. I'm John Furrier, the founder of SiliconANGLE. I'm joined by co-host Dave Vellante, co-founder of wikibon.org. Our next guest is David Scott, SVP and general manager of HP Storage. Previously the CEO of 3PAR, one of my, that was about two years ago. Three. Three, three. David, welcome back to the Cube. It's great to be back. Cube alumni, it was our, our start with HP was three years ago in Barcelona. Amazing. The Cube started in the hotel with that little setup. You, we were there and now look at us. Look, look at us. And now we have a big party. So congratulations on your success. You just gave the keynote up on stage with Meg Whitman. Congratulations. 3PAR has been mentioned. The 3PAR component has been mentioned in the earnings call. Guys are, again, as Dave says, the gift that keeps on giving in terms of financial performance and traction. So give us a quick update. What's happening with the business? And then we'll go into all of it and analyze all why all the success. So give the folks an update of, of the recent business update and all the success. We've been going through a transformation in HP Storage moving from our more traditional platforms to what we call converged storage, which is based on modern architectures, our own IP, 3PAR StoreServe, StoreOnce, our store virtual software defined storage, all part of that. And we're on a tear. Converged storage grew about 47% year over year in our last fiscal quarter that we reported. And that shows tremendous uptake by customers for these modern storage architectures. And overall, we're managing to now handle the transition as some of our older architectures declined. We actually showed a 1% year over year growth. So we're really pleased. We're starting to see the light at the end of the tunnel. So a lot of people say to me, John, this first conversation that they bump into me at the holiday parties is 3PAR a one-trick pony. And I say, no, I think 3PAR is continuing to evolve and more importantly within HP. So that seems to be a question we want to ask. Also the success of 3PAR is obviously well-documented and then the acquisition and the continued success. But what is that enablement? Why are you guys doing so well? Explain some of the dynamics going on around how you guys came in and just created such gravity around the momentum of around you around HP. Well, first I think HP did a fantastic job of the integration of 3PAR. It really made sure that it protected the value that it had acquired. It didn't do anything stupid. It made sure it took the best of 3PAR and actually integrated into HP. So I think we managed to protect a huge percentage of our people, so we kept our core engineering talent. And that has allowed us to, with additional resources, really accelerate our investment decisions. And that has resulted in us taking advantage of this modern 3PAR architecture and delivering it as a polymorphic entity. One that can not just have its traditional high-end implementation, but the new mid-range 7,000 series we introduced a year ago, which is on an absolute tear, and then evolve it into a flash-optimized storage array with a 7450. And people like simplicity and that's what the 3PAR architecture's delivering. You know, I've always been, always jokingly about polymorphic, always joking about it, but you know, so where you created, we found that out last year when you rolled that out. But in a way, it is significant. So let's, I want to unpack that a little bit for the folks out there. Polymorphic is a term that you coined, you know, meaning, you know, single architecture, admit storage. It's right, polymorphic simplicity was the phrase. Well that was the vision, but it was based on a polymorphic architecture. But you didn't invent the word polymorphic, it's all the same. No, well relative to storage. Yeah. It's a well-known biology term. But applying it, from trying to simplify the messaging around what you're doing. So talk about how that scales up and down in functionality because storage is, you guys know the high end. Yeah. So how do you move up and down the performance curve with this polymorphic? Explain to the significance of that, of the movement up and down the performance curve and the interface into that for the customer. Well I think one of the key elements of a polymorphic architecture is that it actually has to start at the high end. It's got to be a multiple controller architecture because then you can bring that down. And if you're forensic enough about the process, as you bring it down scale, you manage to lose none of the critical functionality, but you are able to reach price points or different performance points that can add specific value to different customers. And that's what we've been able to do, not just with 3PAR by the way, but with our store once backup architecture, which we've just refreshed as well. Another polymorphic architecture that can be deployed in many different shapes and sizes to meet different customer needs. So to do that price performance, moving up and down that curve, compared contrast, the alternative. To buy, customers have to buy different, they're with a competitor, different solution points, different interfaces. When you say simplicity, do you mean interface, do you mean? Well if you just take a major competitor EMC for example. Okay, we'll say EMC. Why shouldn't I? Yeah, just say EMC. If you want a tier one architecture from EMC, they'll sell you a VMAX. If you want a mid-range platform, they'll sell you a VNX2. If you want an entry level platform, they'll sell you a VNXE, a different architecture. If you want to go to all flash optimized, they'll now sell you an extreme IO platform. And of course they have to have other architectures like V-Plax and RecoverPoint to make them all work together. Five, six, seven architectures, just for primary storage. What does that mean for the customer? Explain, what kind of pain point does that mean? Well think about five, six, seven architectures, different approaches to training, different approaches to manageability, different approaches to interoperability, replication. It's just more complex, it's harder, it costs more money to operate those environments. So back in the day where there was one app per server, that kind of made sense. And I like interviewing you because you've got an interesting career. You've got a technical background, you've done a lot of stuff in servers. A lot of people might not know this about you. At one point, you were in charge of the monolithic product within HP, the one that you guys resold from. The XP platform. The XP, right. Which at the time, back in the day, having this circle best of breed for each of those opportunities made a lot of sense. And then you left HP and joined 3PAR and so this utility storage thing emerged. You picked that. There were a lot of companies trying to do just that, the whole tier 1.5 thing, a lot of block based stuff. So I want to go back, so what was it that you saw back then that led you to 3PAR and allured you to? And then what's happening now? How would you summarize the big tectonic shifts now and how is HP taking advantage of them? Well I think first of all, back then in the late 90s when I moved into the storage business of HP and we brought out the XP platform, it really was the world-class platform for the type of IT that was being deployed at that time. What we now call a traditional IT approach. But even at that time, you can remember the first generation of application service providers. They called them ASPs. You had some of the storage as a service providers if you remember people like storage networks, et cetera. And it was clear to me that delivering storage and IT in general as a utility service was inexorably going to become the direction that the IT industry as a whole would move in. And so when the opportunity with 3PAR came along, I could clearly see that it was an architecture that could be optimized for this new IT as a service world. And that's why I effectively left HP to go become 3PAR CEO. Roll forward 10 years and that view of the world actually came true. You had the emergence of not utility players, they're actually called club at cloud players, public cloud, some of them are the managed hostess that allow you to deliver a virtualized enterprise applications in a public cloud environment. Some are others are the public cloud providers where you're building a new platform as a service offering people like Amazon, HP Cloud, Rackspace, et cetera. And there was a gap too. Guys like storage networks went out of business before it was a big... Well, the reason why storage networks went out of business is that they tried to have this new utility model using platforms that were designed, platforms like VMAX or really the Symmetric Starter or Starling Apple, just too expensive, inflexible, not agile enough. And that's why we built the 3PAR architecture. It was to be a tier one mission critical platform for the delivery of IT as a service. Now, rolling through our first customers were those people who were in that public cloud, managed hosting service provider marketplace. But we could also see that the enterprise would have to respond to the emergence of the public cloud by building out their own private cloud services to provide the equivalent agility to their lines of business. Otherwise they'd be relegated to less and less relevance. And that's where the synergy with HP came back in because there's a major supplier to large scale enterprises who would want to deploy private clouds. HP then wanted the same technology that these managed hosters were using in the public cloud to deploy to enterprises. And that's why the synergy of the 3PAR acquisition made such great sense to HP. And you've really seen it translate. We went from $190 million worth of product related revenue at time of acquisition to last year we completed over a billion dollars in revenue just from a product perspective, not including services. So it's been a really successful transition for HP. Yeah, and there's no way you would have, I don't think anyway, as an independent company got into a billion nearly this fast. We couldn't do it as fast. I mean, the great thing about HP is it's amazing customer coverage and it's fantastic channel partner network. And we knew that to really blow out the 3PAR architecture on a wide scale, having that marriage would make an awful lot of sense. So let's fast forward because you noticed the sort of service provider trends, whether it was ASP or a storage service provider and obviously that morphed into cloud service. And it seems like they're a harbinger of major trends and forces in the industry. So you see what Google and Amazon and Facebook and Twitter and guys like that are doing. It's not block-based storage. I mean, they have block-based storage, but it's predominantly, initially it was this commodity-based storage, kind of interesting to hear Amazon say they're going to more customized storage. They're like, oh, that'll make David Scott happy, ASICs aren't going to die. But what do you see now? What are you learning from those large internet players and how do you apply that to your business? Yeah, well, you know, the, what I call the public scale out cloud vendors, if you think of the typical, the Amazons, the Azures, et cetera. For them, at the scale that they're operating, it makes real sense for them to use industry standard technology for pretty much everything that they can do within a computing environment. And one of the advantages of being HP is as that evolution takes place between virtualized enterprise hosters on the one hand, people like the services of the world and the kind of public scale out providers, if you like the Amazons of the world. We have great technology for the first with platforms like the three-part store serve, but for the second type of provider, our own ProLiant scale out server technology is a platform that they can use too. So for HP, we win either way, it's just with a different product line. For our competitors, people like EMC and NetApp, they will try and position their solutions for the traditional virtualized enterprise hosters, but they have absolutely no play in the public cloud. And that's an exposure that they have that HP doesn't. Well, they would say their play is the software-defined play as that evolves. And you know, they're smart people and they see the industry trends too. So let's talk a little bit about software-defined. So again, the big public cloud guys have always done that software-defined thing, sort of programmable infrastructure. And maybe you didn't catch my new one. So I should have been a little bit more clear. At AWS re-invent, I was learned in talking to some of the Amazon architects that they've done a 180 on infrastructure. It's not just commodity-based infrastructure anymore. They're pushing suppliers to do very much highly integrated, customized hardware, which is, I think, I'm not surprised, frankly. We've heard that the hardware for the last 25 years, maybe more, and it just, value keeps coming at the hardware, but nonetheless what has changed is the ability to do software-defined without sacrificing performance. So what do you see in terms of that software-defined meme and how does that affect HP? Well, I think in the area of software-defined data sensors as a whole, HP has an incredibly strong position. We have software-defined networking with controllers that Bethany Mayer, my colleague, has deployed. But we've also had a leadership capability in what is called software-defined storage in the form of being able to run virtual storage appliances on industry-standard hardware and being able to convert that into shared resilient storage infrastructure by clustering those industry-standard servers together. So we have that platform delivering already today. And in fact, it's been successfully delivered since our left-hand acquisition as our store virtual PSAs for almost five years now and have 175,000 of them deployed already worldwide. So we're the clear leader, very rich functionality. Now for any second or third tier service provider that wants to build a scale-out architecture on industry-standard servers, and they don't have 50 or 200 PhDs to be able to write their own software layers, this is a really excellent alternative. But by the way, it's also a great alternative for a small, medium-sized business or remote offices and branch offices as well. But turning back kind of the scale-out world, we fundamentally have this leadership technology. We've now extended it into the backup space. We brought out at Last Discovery in Las Vegas to store once VSA. And so we think we have real product that is battle-hardened proven that we can deploy to a wide class of service providers as well as these small, medium-sized businesses and remote offices with software-defined storage. So we're not scared of vendors like EMC or VMware coming into this market. And we also have the huge strength, Dave, of our own industry-standard server business. What you may not have caught is that about four weeks ago, we announced a program that on every eligible ProLiant server and even our blades, we will ship a free one-terabyte virtual storage appliance license, store virtual. That means that a customer can take three ProLiants, cluster them together, run the VSAs on each of the three servers and have a three-terabyte resilient shared storage platform and then run their application side-by-side on the same physical infrastructure. 60% lower cost, much easier to operate, no skill sets needed. And we think that technology is going to be also applicable in many of the service providers, tier two, tier three service providers as well. Set against that, what does EMC have? Viper. Viper, it's a management layer. It's not really software-defined storage. They actually went out and bought Scale.io as a platform, still not GA, some point in the future, an unproven technology will be brought to market like Extreme.io. VMware, they have their VSAN technology, which they're planning to launch, they've announced, but it's still not GA. So we have this leadership position and we feel very good about it and we're now able to leverage this massive, we sell millions of units of ProLiants every year. Each one of those represents a potential upgrade opportunity for more store virtual VSA licenses as we allow these customers just to easily upgrade their existing infrastructure. The moonshot for cloud service providers is kind of interesting as well. We'll see what the uptake there looks like, but, and I agree with you Viper, it's very early days, I'm not even sure exactly what it is yet. Well it's another incarnation of storage management software as I can see. Yeah, so there's another layer there, but the concept, forget Viper for a second, but the concept of having a restful API whereby you can make a call and manage your storage, whether it's block or file or object is alluring. Do you see that day coming and also function, like for example, function that the secret sauce inside the three par ASIC, do you ever see that going into software defined and maybe being replaced by some other secret sauce function, what's your outlook there? So first of all, let me outline the broad view of how software defined connects to cloud and connects to virtualized infrastructure because if you look at virtualized infrastructure, our converged infrastructure strategy at HP, it's really all about the administrator having control. If you look at the cloud, the cloud is all about the end user having control through self-service portals. Or the developer. Or the developer. The difference with software defined data center is now you're able to hand control to the application or some kind of restful web services API. So you now have the potential admin control, user control and application control. And we feel that we're in a very, very strong position as an infrastructure supplier in this evolution that includes the software defined data center. Why? Because to have control dynamically, you have to be able to monitor the infrastructure. You have to take information about the infrastructure, how well is my application performing and then translate that into dynamic policy change of service levels and changing the configurations of the infrastructure that is servicing that application. You need to have the control of the core converged infrastructure itself and that is what HP has in service storage and networking and really uniquely positions us to be able to provide huge differentiation in the software defined data center space. Now as far as kind of taking platforms like 3PAR and moving it into software defined storage, I'm actually a great believer in horses for courses as we say in England. The right racehorse for the right course. And fundamentally 3PAR has always been built around an ASIC structure. If we change 3PAR to kind of use a different approach without an ASIC, it would make it far easier to deploy that technology as a software defined storage kind of implementation. But I have no interest in doing so, certainly in the short to medium term, given that I have the world's leading software defined storage platform with StoreVirtual and our VSA technology. And so our intention is to leverage StoreVirtual where it's really designed and then integrate it really tightly with 3PAR StoreServe that may be working in customers and enterprise data centers whilst they communicate out with their remote office. Well, yeah, that's why you're getting unique advantage for your customers and relative to your competitors with the 3PAR architecture. So until that change or it appears it going to change, there's no reason to mess with that. I understand. Just want to see how we're doing on time because I could go a long time with you. Before we get into, if we have time, I would like to talk about what you're seeing in trends and flash, but I want to get a quick business update from you. Since the 3PAR acquisition, 3PAR has been for a couple of years anyway, was growing at triple digit growth. I think it's now moderated to very high double digit growth. I wonder if we could share what's going on there, the converged storage figures and momentum. Maybe talk about some of the growth parts of your business. Yeah, I think IDC just released their calendar Q3, worldwide external disk quarterly tracker figures. And one of the things that we were really pleased to see in the mid-range fiber channel segment, the largest market, is that by product family, they reported that the 3PAR platform had grown 261% year over year, just an astonishing growth rate. And this is really in response to us creating a revolution in the market a year ago. By bringing out the 7000 series, we brought tier one mission critical functionality down to mid-range price points and affordability. And that has been taken up. Our channel partners love the platform. We're not only replacing EVAs, but also going out and taking out EMC, install base, NetApp install base, and many other papers. So we feel really encouraged by that momentum. In the summer, we went into the all flash optimized array. We brought out the 7450, 550,000 IOPS at less than 0.7 milliseconds. I'm really excited personally that just six months later, we announced, have announced with a pure software upgrade, a 60% improvement in performance, up to 900,000 IOPS at less than 0.7 milliseconds with a very scalable platform. You can support 220 terabytes of all flash in a 7450 now. And that platform matches up with the specs of pretty much any all flash startup that has started from the ground up. But the difference is, Dave, it has full mission critical capability, all of the service, it has the stack. And so it's an unbeatable value proposition. And so we think we're going to see really strong momentum in that market, especially as we added new MLC solid state support that has halved our cost per terabyte. So it's made it very much more cost competitive as well. So, okay, so while we're on flash, let me get your take on what's happening in the flash market. You got hybrids, you got all flash arrays. You chose not to go out and buy a flash company. Others have, I mean, IBM, EMC in particular, why didn't you feel the need to do that? And what gives you confidence that you can compete going forward? Well, it's very simple. When you're looking at optimizing for all flash, you're worried about three things. Performance, endurance, and efficiency. And so when we scan the three power architecture, we realized because it was a modern architecture that was originally designed, if you think about it, for performance, efficiency. And as a byproduct of the way we approached efficiency with system-wide striping and system-wide sparing, it happened to align itself really well to endurance, characteristics of an all-flash array. We didn't have to buy a new architecture. Instead, we could just take the three power architecture, make very simple engineering changes to it, to optimize it for the all-flash world. And so the benefit of having a robust proven tier one architecture apply to all-flash optimized is obviously of significant advantage. And if we can match the performance and efficiency specs without having to go to a new architecture, that is a great investment protection message for our customers because it means that they can turn around and keep their three power investment and then just move forward to the all-flash optimized world with three power. So I wonder if you could clarify something for me because the perception is, of course, it's probably a fact, that a lot of the stuff that the three power engineers were doing in the early 2000s was designed to minimize the latencies of mechanical disk. But you're saying the architecture applies very well from an endurance standpoint to flash. Can you just add a little color to that and explain that a little bit? Well, what we did with our architecture is develop system-wide striping in combination with chunklets. And effectively, that meant that any single volume that we provisioned on a system would be spread over every drive on that system. The advantage of that at the time was tremendous performance, but it was also combined then with a paralyzed clustered architecture of up to eight controllers so you could have multiple access through controllers to system-wide striped systems for great performance, but it also provided with the chunklet architecture system-wide sparing. So the system-wide striping and the system-wide sparing reduce the level of impact on the drive technology. Now, if that drive technology moves from hard disk drives to flash-based, you have a perfect situation for flash because you're minimizing the wear and tear on each individual flash drive. Now, we've actually extended that because we went with our flash drive vendors and looked at it and said, inside a flash drive, you typically have a whole load of capacity that the vendors reserve to spare out themselves in the failed regions of memory. And what we said is, look, we already have a, if you like, a chunklet-based sparing mechanism. Why don't we create something called adaptive sparing where we'll take over that responsibility and then you can release more capacity so our customers get better value. They get more capacity out of a single drive than they do with other vendors operating system implementations. Ironically, it was fundamental, you're saying, to the flash architecture, not just the spinning disk architecture. Okay, let's talk about some of the other products that you guys announced this week. What's exciting you? I mean, we saw your keynote, we've talked about the three-power-all flash stuff, but you've got StoreOnce, Refresh, and StoreAll, Refresh. Maybe talk about those briefly. Yeah, I think with StoreOnce, we really are blowing away the first-generation deduplication-based backup system architectures like EMC Data Domain. They're all unicontroller architectures, two years after we introduced our high-end B6200 with high availability, EMC has still been unable to respond, providing high availability for their backup solution. And what we're doing is steadily extending our speed and differentiation with the latest high-end 6500. We back up 40% faster and we doubled our recovery speed, but competitively, that means that we back up four times faster than EMC's flagship Data Domain 990 and recover an astonishing 10 times faster than they do. Literally, we can recover as much in a single day as it would take them 10 days. Now, there was some research done a while ago with the National Archives that if companies go beyond 10 days in terms of backup, half of them go out of business within a few quarters after that. So it really matters, especially as the amount of data that these companies are storing is exploding. So we have a modern second generation Store 1's architecture, which is really adapted to this explosion in human information, structured information, machine-generated information that customers are experiencing today. So we feel really good about that. And by the way, the Store 1's product line in the fourth quarter of last fiscal year grew 70% year over year. So it's on a tear as well. And then finally, Store All was our information retention platform. Again, just like with Store Once, we revamped and refreshed the whole product line with Store All, the new 8,000 series. It could be deployed as a gateway in front of 3PAR or with very cheap and deep storage capacity in its own right. Low-cost economic storage for you to use as an information retention platform or an archive platform, differentiated by its amazingly fast search speed. When you're on scale of, you know, 500 million files or objects, it can search at 100,000 times faster. And what we've done is simplified the searching. You can now search on multiple attributes. In the next generation of storage, we've Store All, we have increased the density by 150%, 40 terabytes per single U of rack space. And the final thing we did with Store All, which excites me as part of our hybrid cloud strategy, is that we added this open stack interface to it for Swift and Keystone, their object and identity service offerings. And that capability means that now customers can develop in the HP cloud or at rack space, you know, or anyone using open stack, and then deploy safely within their own enterprises if they're hesitant about, for security reasons, about deploying out in the public cloud. And get that Store All integration out of the box. Absolutely. All right, good, we get to the hook, but tomorrow you have a deep dive breakout session for people who want more information from your keynote today. That's 9 a.m. 9 a.m., Hall 8, Romeo 02, I think, A02. Well, David, always a pleasure. Great place to start off the day. Yeah, why not? All right. Thank you, Dave. Thank you very much. John and Sharon, your insight says always, keep it right there, buddy. Like, we'll be right back. John, Ferry and I, with our next guest right after this. This is theCUBE.