 Welcome to the Cube. This is the first for us, John. Inside the Cube. Intel. We're inside a lot of stuff, but I don't think we've ever been inside the Cube. Intel inside. We tried to get Intel in the Cube multiple times. They have not returned our phone calls. Former Intel. Pat Gelsinger's been on three times. Oh, there you go. Well, Pat is a mentor of mine. He says he does an Intel bong before he goes to sleep every night. Is that right? We have that live. You have? Pat doing the bong? No, just talking about doing it. Oh, OK. I would pay money to hear Pat do the bong. John said it probably helps him sleep at night. Yeah, there you go. So what is the Intel data center group? Share with our audience. Well, Intel is organized into several different business groups. And recently, about less than a year ago, we actually formed the data center group, which is a merge of our server platforms group, where Intel has been working on Xeon processors in the server platform for over a decade. And our storage group, which is what I run for my boss, Kirk Scalgan, who runs the data center group, and our networking logic group. And all three of those divisions form the data center group. And it's because we believe that, and you can sort of see it from the VSP announcement today, that storage, networking, and compute, or servers, are actually becoming very converged in both architecture and in different ways of being used. And everything's being just sort of distributed and cloudy. And there's also a lot of innovation, as you can see by some of these really ridiculously awesome mergers going on around with great prices, right? Yeah, so you guys obviously just recently had a big acquisition, software acquisition. But it's known in Silicon Valley that you guys have been high on the whole software bandwagon for some time. And you guys are known for the chips. So talk about what's going on with the software side of the business on the integration, because you guys are bundling software into the hardware. How far has that come and where are you guys today? And what is that for your IT market? Yeah, well, I mean, software, not done, but the McAfee announcement is one of our biggest, is obviously a growing vector for the Intel corporation. We're still a building block supplier in the end on the silicon side and on the solution side, so we make chips primarily. But as we've been looking at growth, it turns out that there's quite a lot of value we think we can bring by adding certain key features in software that goes with those chips, kind of extending that software with silicon features. And that, I think, is an opportunity for us to add more value. But we'll do it in a way that's very open. We have all sorts of industry standards we drive too. And so you'll see us anchored on delivering great chips, and the architecture goes with that. We've been building driver software, and like we have a RAID software that goes with our upcoming chips for servers. Those types of building block softwares are tightly linked to our chips. And now, of course, we have some other areas, and the security which we think is a huge area of growth is one of the areas you'll see us focus on in addition to manageability. Manageability, security, and energy efficient performance nowadays is really a big part of Intel. So other than Moore's Law, what's driving your business? And again, how does that affect your IT? Yeah, well, I mean, Moore's Law is a religion with us. It's our culture. It's really hard, by the way, to make happen to double those transistors every two years. Thank you. It's actually quite, it's kind of not easy. But what's driving, like, I'll give you an example of what's driving my business today, is this explosion of digital data. You know, Facebook and, you know, your show and everything, you know, is a big deal for online. And I mean, look at the IDC report that just came out called the Digital Universe Study, and every year they come out with this and every year they up their numbers. But last year in May, they said there were almost 800 exabytes. Of data, digital data out there, which is a 10 with 18 zeros after it. We're talking zettabytes these days. It's unbelievable. Yeah, and I didn't know what a zettabyte was until recently, right? And so now, you know, they think it's going to grow 44x, which is actually faster than Moore's Law. And by 2020 and 10 years from now, they think it'll be 35 zettabytes. So that's definitely going to grow our business. It's a big deal. I mean, we obviously think our show's a big deal. Thanks for pointing that out. And we run out of storage all the time. We do. There's a 1080p video. We're putting out monster files. I mean, this is like, I can't believe how much storage is required. But the question is, is that there's a big debate for the people who aren't in the business around understanding what's going on with virtualization. And when Paul Moritz comes out, he quotes the IDC study and says, there's more servers being virtual, a lot of shipments. More virtual shipments in servers than actually physical. The average person says, oh, that's bad for Intel. So clarify that, because we've said publicly, that's a great stat and gets people juiced up. But really, the data's not going away. The machine footprint is still continuing to grow. But just clarify that fallacy or the artificial debate. Yeah, yeah. Actually, we're big fans of virtualization. And we built actual core technologies into our processors and into our chip sets to virtualize both the processor itself and help it run faster and better and into the IO. And from our perspective, it's just, you know, even though we love Moore's Law, we love to solve everybody a processor per usage model, it actually makes a lot of sense. And so if you can consolidate your server infrastructure and run several virtual machines onto a device, you get great performance and you save a lot of power. And for us, you know, we think of it as, well, we'll just sell a better processor, right? So instead of selling a simpler processor, people tend to buy our dual processing or our four processing, our quad sockets. You know, not quad core, but quad sockets. 16-core, 18-core, 20-core solutions to run these, you know, Oracle databases, SAP, all these very high-end things. Now, you know, virtualization does have some limits right now. You know, it's really being used for what some people call tier two, some sort of cast of build called crapplications. But, you know, tier two, you know, apps, simple apps, and we think that it's going to have some penetration limits. So you might run five virtual machines, you know, on a server, and you'll run it for your tier two apps. But the high-performance stuff, you know, the databases, it seems like they still want, you know, very, very direct access, right? So as you put Jim, more cores into the equation and maybe Flash also comes into play, what does that mean for I.O. and connect it to what you're seeing today with this Hitachi VSP? Talk about that a little bit. Yeah, well, so, I mean, the key thing about Flash in a storage application is data caching and some people call it tiering. And so today, before the world of NAND and non-volta storage, the fastest way you could get access to the disk to data was to concatenate hundreds of disks and put the data literally on the outside of those disks so that when it rotates around, you get the highest centrifugal motion. That was speed, right? And obviously, you know, you're talking about 200 I.O. ops on a disk. This is Intel GeekySpeak, I apologize. But, you know, you put an SSD out there and you get 10,000, 20,000, 30,000 I.O. ops. It's just, you know, ridiculously different and better. And so it's being driven into the data center as a performance tier anchored on transactional database things. But, you know, once you have it, it's useful for a lot of other things. And you can see from the Hitachi VSP announcement today, it's sort of integral to the architecture, which we think is essential because it's hard for an IT person to sort of on-the-fly dynamically manage all this tiering. And it's really better if the storage box sort of owns that problem and sort of abstracts them away from it and does it logically based on their knowledge of the data. But we expect non-volta storage to be driven, you know, very, very steeply into a sort of external storage box, but into the data center in general. So you're a USPV current customer, is that right? Or a current USPV customer, a Hitachi customer of their previous platform? I'm sorry, I don't understand the PV. The predecessor? Oh, the predecessor? I actually don't know. Yeah, okay. I actually don't know. So Intel, so that makes a lot of sense. So just summarize the question for the folks out there. I said Intel actually is interested in virtualization, enabling virtualization because you sell more high-end processors and more stuff. Yeah. And just to tie back your earlier point, you guys have IP, specific IP in the silicon and you're going to wrap your software into that. You see that as that major. Yeah, yeah. And, you know, the data center's exploding, you know, and we're already in 90-ish percent of the servers run on Xeon for the target segments we identify. But you may not know this, but actually by the end of this year we're probably going to be 70% of the storage boxes out there. We'll be running on Intel processors. Who's left that doesn't run on Intel? Well, you try to make everybody happy. Even Oracle's running on Intel. So kind of the wild card question is, people are talking about this, is that people are building their own. Facebook is, you were talking about Facebook. And Facebook's making all their own stuff. Google makes their own machines. Facebook granted software they're building. But there's a trend to, you know, hey, Intel who? AMD, who's got the better lower power, you know, better everything. So talk about that opportunity and what's happening around that. Is it homebrew-like stuff? Or is it real? Yeah. Where's the market at on that? Well, you know, yes. So if you had shown my top 10 customer lists in the data center group five years ago or today, you wouldn't see guys. Now for us, we sell the chips and we sell them into OEM platforms. So this is HP and Dell and IBM and lots and lots of other companies who are building our products and then selling them off to the Facebooks and the Google's. We don't necessarily do chip directly to them. As a business model. But, you know, that being said, they tend to want to strip it down and do a lot of their own stuff in software. So they might buy just a basic server and then do a lot of cool things in software with it. But they'll buy a ton of them. And matter of fact, our top 10 list of customers includes if we measure it by consumption and measure includes Amazon and Google and these types of fuss that are in our top 10 and they weren't five years ago, 10 years ago. They didn't exist, right? The purpose built is a big message from Larry Allison when we're all this notion of integrating hardware and software. And as he said, my friend Steve Jobs referring to Steve's with Apple strategy. But Oracle's more of an incumbent. But is that the trend more appliances, more purpose built hardware? Or is it just a fad or a specific? Well, I don't know. I think it's certainly a trend to keep trying to take particular bottlenecks and pain points and there's just constantly new bottlenecks and pain points in the data center and build accelerators to solve them. But actually I think the thing that we're seeing is what people call convergence, right? Where the server, the networking and the storage boxes are concatenated together in their common management solution for them and they're even sometimes sold as a module. Pretested. Pretested, configured, whatever. And so that convergence, we think, is sort of logical for IT people because they touch it one time. But we also think that people want to have best of breed, right? So they want to be able to interoperate all these things, right? And people select to whoever they want for that convergence. And so I think we, being a chip supplier, our goal is to make sure architecture works great in any of those environments. And it's quite a deal. I mean, our Xeon chips are very different from our client chips, right? And of course we have the Atoms for the low end, right? And so we spend a lot of time and energy trying to figure out how to get specific technologies to these different boxes together. And Oracle is a great partner of ours. They just announced their new storage solution is using Intel architecture. Yeah, no spark inside. And so we consider that to be a good thing. But I think that's where I think you'll see it happen. You'll see, yes, definitely appliances, but I just think you'll also see this kind of converged platforms wrapped by an amazing amount of software. This digital footprint idea you mentioned earlier is pretty big. I mean the storage of digital media, content being one of them. It's mind-boggling to think that my Facebook photos will be there forever. And we have a historical record of digital content. Will that be the case, do you think? Or will there be a trash pile of recycled components? Or will the photos be around forever? It's a little outside my expertise, frankly. I don't think I've ever really studied it. I think the world's been so focused on trying to keep up with the explosion of the data right now, frankly. No one that I know of is having a top of mind conversation about archiving and what you should get rid of. Like right now I'm just trying to run to keep up with the number of pictures being created. I took my son to a university tour. He grabbed my phone and he took a shot of the school mascot in front of him. And he had it posted to Facebook before I even got my phone back for him. Right? Thanks, Dad. It was amazing, right? And that's the type of stuff that's going on. So I don't know. I think it's going to be a lot of everything, and who knows how it's going to turn out. Good. Well, we're here with David Tuey, who's the general manager of the Intel data center group, talking about crystal ball, futures, current... Great to have you on theCUBE. You're the first Intel person on theCUBE. We've been trying to get someone from Intel for a couple of months now. Well, we won't hide from you, I guess. We're not... Oh, they're right down the street. I live in Palo Alto. Oh, yeah. Well, the world is coalescing around big waves like virtualization and Intel's at the center of that. So it's great to have you on. Appreciate you sharing your thoughts with us. Yeah, thank you. I have one more question for you. Well, I don't even know what a dark cloud looks like. But I'll tell you what we're doing with security. And we kind of feel like it's something we should take very seriously, because it's a worldwide concern, right? And people want their privacy and security. And so what we've been building, for example, into the Xeon processors is something we call Trusted Execution Technology, which is an ability to securely boot to a known virtual machine or OS, the PC. And you've probably seen some white papers on it. RSA and Archer announced a white paper back in January. Right. And the idea is that, you know, would you rather have a, frankly, it's to be simple, a secure cloud or an insecure cloud, right? It seems like a simple question to answer, right? But that core technology is, when you boot the OS, you don't want anybody in the middle between when you get that system started, when you get that OS up and running, and it takes control. And so Trusted Execution Technology is already a technology, and we've been investing in that for several years, actually. And now it's the software system that's starting to catch up to it. But those are the times of things we're trying to do. We don't put crypto in our chips. We put, like, accelerators and building blocks. But we do put security features in there. And we will continue to do that because, you know, the attacks are, you know, they're big, they're escalating, and we all know, right? And, you know, they're subversive sometimes. And so we feel like it's something that we should offer. A lot of bad guys out there.