 Live from Las Vegas, Nevada, it's the Cube at HP Discover 2014, brought to you by HP. Good morning from Las Vegas everybody, this is the Cube and my name is Dave Vellante and I'm here with Jeff Frick. This is day two of HP Discover, we actually came in early, we were meeting with some customers of HP's on Monday night, unpacking some of their secrets. Day two yesterday was a big day, we had David Scott on, talked in depth in storage, had a bunch of customers on. Today is really around a combination of, we're going to be interviewing customers, we're going to be hearing from Bill Vecti, who leads the enterprise group. Bill Vecti took over for Dave Donatelli, who came from EMC and has now left HP. So Vecti is now the man in charge of the enterprise group which comprises servers, storage, networking and technology services. So we're going to hear from him, Bill Vecti was the COO of HP and helped architect HP's cloud strategy, so he's a super senior executive, one of Meg's right hand men and somebody that we're very excited to have on the Cube. As well we're going to hear from a lot of the server folks today. So you're going to hear from people around what HP announced on something called Apollo, which is a big supercomputer, high performance computing capability, and others from the ISS group, the x86 business. So we're going to be unpacking a lot of that today. We've also got some fun this afternoon. What time do the puppets come up? 12.20, we'll be joined by Mario and Fatah, the groundhog from Gloven Boots. If you've been paying any attention to the show at all, you can't have this system, they did a funny like a six minute commercial, they were at the party last night, they actually stopped by the Cube yesterday to get the lay of the land and we posted a couple of pictures up on the Twitter feed. So they're coming in and also we're getting more customers, Dave, which is kind of a nice trend here at HP. A lot of times, well, as a matter of fact, we never get enough customers, we always want more customers. And so we've got a bunch coming in, looks like we've got the healthcare industry represented pretty well. And in big cloud day today, so Kerry Bailey is coming on, he's the new senior vice president in the go-to-market side of things, Saar Goli is coming on, Saar is the COO of HP's cloud, and essentially really architected HP's cloud strategy under the direction of Bill Vecti. And so HP announced Helion, I always say that wrong, it's Helion, it's like Helium, Helion. They announced Helion a few weeks ago and it is really there, it's kind of there, I would say second or third generation of cloud, they announced the HP public cloud a while back, which was frankly largely vapor, but they've now started to bake it out. And so it's really about the hybrid and that's what we're going to hear from those guys today. So big cloud day, John Furrier flew back to San Francisco for another event that we're doing there, so you'll see some of that action going on around Flash. And that was the big theme yesterday, I want to come back and talk about that, and then John will be back, I'm sure he's excited about the cloud action. And we're also going to hear from the Moonshot guys. So Moonshot was a product that was announced a couple of years ago, the new type of server, very efficient, highly dense, using cartridges to essentially change the personality of the server. It's something that HP.com has run on, HP.com gets a lot of hits. We want to get an update on that, you know, see where it's at. Haven't heard much from the Moonshot guys lately, and they are going to be on the cube to give us an update there. So Jeff, you, so last night, yeah, we were at the storage party, Dave Scott, David Scott, and was sponsored by Emulex, Brocade, and QLogic. It's always a big, big bash, it's at excess, it was at the encore. We stopped by, we kind of did a, did an early evening because we were prepping for the cube today. But the big thing, the big news yesterday was HP announced an all-flash array last summer, and at the time, we said, okay, looks good, it's got a full stack. Now for those of you who don't follow this market closely, it takes a long time to build up a full storage stack. Specifically we're talking about things like snapshots, thin provisioning, resiliency, data migration, all the storage services that guys like EMC have built up and perfected in NetApp over the years. I mean, decades to build this stuff out, really robust, mission critical, you know, high-end capabilities, mainframe-like capabilities for, you know, mid-range costs. So 3-Power had that, 3-Power said, okay, we don't need to go out and buy a new company to compete in the all-flash array like IBM did with Texas Memory Systems, like EMC did with Extreme I.O., et cetera. We have an architecture that we can adapt for flash. When they announced last summer, a lot of us in the analyst community felt it was good strategy, but too expensive. They had more expensive flash. They didn't have data deduplication, like the guys, for instance, from Pure, that are really pushing data dedupe and compression. And it just felt like it was early and expensive. It was about $13 a gigabyte. Well, they announced yesterday that they are using lower-cost flash, and they are applying data deduplication in line to the product, and they've got prices down now below $2 per gigabyte, which is comparable to high-speed spinning disks. So this is the day we've been waiting for, the day that flash replaces high-end spinning disk, and eventually, you know, it'll eat into the core, the disk market, that will take some time. Like our laptops. I don't know about you, but I don't use spinning disks in my laptop. You don't have spinning disks in your laptop. Only thing we use spinning disk for is I carry one around in my bag to back up every now and then. It's my archive. So that's what's happening in the enterprise. The spinning disk is becoming the bit bucket. Yeah, you know, and we like sports, we talk a little about sports, and I always think of there was a great analogy with NFL films back in the day when shooting in slow motion was a lot more expensive than shooting in regular film, and they asked, they asked NFL film, so how did you get that play in slow motion? How did you know that that was the one of the hundred that you shoot in slow motion? And he said, you know, we just shoot everything in slow motion because we don't know when that play is going to be. And I just get this feeling with all the vibe around flash that the adoption of flash, the benefits of flash-optimized applications is going to push that a lot quicker than I think most people are giving credit to. And then, you know, a lot of talk here at the show about this magic price point. The $2 per gig price point, which sounds like we've achieved and is a critical tipping point. You've covered the space forever. Is that a big tipping point? Here's what's happening. I think it is a tipping point. And the tipping point is not so much the cost. I mean, it is, it happens to be that in June of 2014, that happens to be the price. The tipping point really is that flash pricing is now getting close enough to the comparable high end of the spinning disc market. And that is the tipping point. Now, I got to give props to David Floyer. I've been working with David Floyer for a number of years and he's made a lot of good calls over the years. One that he made back in 2009, he did a flash. You can go in the Wiki and check it out. Flash, cost, and pricing forecast. And in that, he said, well, there's a lot of innovation coming in flash. And prices are going to be coming down between 50 and 70% per year. So he ran a model and he said, here's what's going to happen if it goes, if the pricing comes down on a compound basis. 50% a year, here's 60% a year, here's 70% a year. So he used 60% as sort of the mid case because that's what he thought it was going to come down at. And that's essentially what it was tracking at up until recently. And in 2009, Floyer projected that by 2014, flash would be close enough to high end spinning disc. By high end spinning disc, we mean high spin speed fiber channel and SAS disc spinning at 15,000 RPM, so-called high performance disc. That by 2014, the prices would be close enough that essentially it would be the tipping point. So that really is the factor. And I'm just so impressed because Floyer was, he was right on it. I mean, it's June of 2014. We're right in the middle of the year and it's exactly what he predicted. So kudos to him. Now what's happening is prices have been coming down for a number of reasons. One is the use of lower cost flash. The second is the higher capacity flash devices. And the third is the use of technologies like compression and data de-duplication, which you can use on flash. It was harder to use on spinning disc because it would slow the spinning disc down even further. And that's the last thing you want to do with spinning disc. So for instance, NetApp popularized things like data de-duplication and compression within their arrays. They would give it away for free. But many customers didn't turn it on because it hurt performance. The only company that has a compression that really doesn't fundamentally hurt performance of spinning disc is IBM, with IBM real-time compression. And they're finally getting that to market. Now the point is you can use compression and de-dupe in flash because it's so fast if you implement it correctly. And that's what's happening in the marketplace. So all the innovation in flash is causing flash prices to come down faster than disc prices. And that is really causing the tipping point. Well, the other big thing, Dave, and everything you just went through is all about cost, cost, cost, and delivering on the same applications with different infrastructure to drive down costs. But what you didn't talk as much about it, we talked a lot at actually Hadoop Summit last week, is really the value that these new applications are unleashing with the internet of things and big data that now, you know, marginal increases in the performance of these applications or slight variations in the performance of these applications or enabling the speed of business to move faster with big data and these flash-enabled applications so people are making decisions in real-time. The business impact at scale is huge. So, you know, we talk about cost-based pricing and value-based pricing, and it really feels to me that this flash on this complete, perfect storm of big data in cloud and infrastructure and like I said, Moore's law applying across the entire stack is now making that flash investment look a lot different from an ROI point of view as opposed to simply trying to replace existing infrastructure. Well, let me comment on that because you are right on. That really is the exciting part about flash. It's not all, it's the cost, it's the value. Now, having said that, it's all about economics, but the reason why we're so excited about flash is because spinning disk, I like in spinning disk to a military convoy, I've used this analogy many times where the slowest vehicle in the con, all the convoy has to slow up, all the vehicles in the convoy have to slow down to allow the slowest vehicle to keep up. That's what disk drives are like. The spinning mechanical disk is the bottleneck in computer systems. So, flash eliminates that, but not only. So, what you're seeing, and you're seeing advancements from guys like FusionIO who have a strong relationship with HP, they're tightening that relationship. FusionIO now is run by a bunch of XHP guys, particularly Shane Robinson. They just had their big announcement the other night. Yeah, you guys were there. That's right. The Atomic series launched, which really, if you read, again, Floyder's research piece on the benefits of atomic rights riding once and not twice, on the benefits of the MPM compression, on all the benefits that are basically enabled in the software, sitting on top of the really commodity flash component, and that's really what they announced in this Atomic series. So, let's unpack that a little bit. So, Flash 1.0 was stuffing flash drives into essentially a VMAX, EMC announced it, and it dealt with some of the performance problems that they had, but it very quickly ran out of gas because the controller bandwidth wasn't designed to accommodate flash. So then you saw companies come out with, actually at the same time you saw FusionIO come out and say, let's put the flash right into a PCI-E card, closer to the server. So, you had them at both ends of the spectrum, at the end of the wire with the disk subsystem, and as close to the server as possible. And Flash has been filling in that entire spectrum with hybrid arrays like guys from Nimble who were covering today at their event, to all flash arrays, companies like Pure and Violin, and now ExtremeIO and HP and many, many others. Everybody's working on this stuff. What you're seeing is not only is Flash eliminating the spinning disk mechanical movement, but with technologies like atomic rights, which FusionIO is really one of the few vendors, if it's not the only vendor shipping today, EMC made the acquisition of DSSD, and that's I think what this is all about. But atomic rights essentially directly right to Flash as an extension of memory. So think of it as an expanded memory, with a direct memory access protocol essentially, with primitives that allow you to write directly to the memory. What that does is it eliminates the SCSI protocol overhead. So SCSI is the disk protocol which has been hardened over the years. It's very chatty. It's not conducive to performance, but that was okay because when you spin the disk, it's so slow anyway, so you add a little bit of overhead at no big deal. Well, when you eliminate that spinning disk, that SCSI protocol becomes the vast majority of the bottleneck. So atomic rights essentially eliminate that. So you're going to see though a full Flash spectrum emerging, it is emerging, and it's very exciting for the following reasons that you brought up earlier. It changes the way applications are going to be designed. An application programmer realizes that when he or she has to go to disk and do a write, it's going to be slow, so they have to go off and do other things in parallel. If you have a virtually unlimited resource that's persistent in memory, you can now change things. You can do writes directly to that memory and change the way in which you design applications and achieve not an order of magnitude, but potentially two orders of magnitude performance. That's exciting. That's going to support mobile. That's going to support pushing data to the edge. That's going to support video applications and all the new emerging applications that we're seeing out there. So that's why we're so excited about this. Yeah, and applications built specifically to take advantage of this capability. So it begs the question, and what's the next point of failure, right? Every, as you mentioned, every process you're always gated by your slowest piece of the chain. Well, I think it's a network now. I think it's a network. Yeah, so that's why everybody's, so networks are very hierarchical. I said yesterday on theCUBE, the networks in the enterprise are like the Kremlin, very hierarchical and structured, and what's happening is those networks are flattening. Traffic used to be the emphasis was on North-South traffic, and now it's about East-West getting stuff out to the edge, wireless networks at the edge, supporting mobile applications. So I think that's where you're going to see some of the bottleneck emerge, and so that's why people are so focused on things like software-defined, improving flexibility, and obviously higher performance. Two other guests that I want to mention today that are coming up, Robert Young-Johns is taking over, has taken over for George Kadifa, whom you know quite well, and the Palo Alto guy, is now running the software business. He is the Senior Vice President and General Manager of HP Software. In fact- He hit his keynote yesterday with a ton of steam. Of all the keynotes that we saw yesterday, he was really charged up and really came out guns blazing. And then Antonio Neary who runs the server and communication and networking business. So Antonio used to run the services business and now runs the networking and server business. We're going to hear from him. We got to leave it there. We'll be back with our next guest right after this. This is theCUBE, we're live from HP Discover. This is day two, Dave Vellante and Jeff Frick. We'll be right back.