 here with George Simons, who's the COO of Xiotek. As we all know, storage is a major pain in desktop virtualization and VDI. Storage is always a major pain, isn't it? It is. So you have to manage it and always challenge it. Xiotek's a pretty cool company. You guys have a radically different philosophy around storage, which essentially, you're trying to simplify the concept and let the application do its job. But why is that important for VDI? What's the fit with your philosophy in VDI or desktop virtualization? So there are two fits. The first one is virtualization in general. And VDI in particular requires a tremendous amount of performance, IOPS. And really, that is the core of the ISE technology, is to deliver performance out of really the same technology that everybody else has from a drive. ISE, meaning the intelligent storage element? Very good. A brick. A brick. It is a brick, and it has sealed data packs. I think you had one of our guys on a couple nights ago demonstrating and showing you what that sealed data pack is. So the whole idea is you put this in, number one, you never have to touch it. You don't swap drives. You don't have to pay attention to it. It just runs. So that's the second side of it. It just runs. But when you get to VDI, it really is about performance. And the fact that we can get twice the performance out of the same spinning disk is really important, because storage is the bottleneck in two ways. Yes, performance, but cost. If the cost of a virtual desktop implementation is too expensive, people aren't going to do it. Why replace my desktop with an expensive virtual desktop? And it's because people have had to over-provision the storage. You have to put more spindles out there to get the performance. We provide that same performance without over-provisioning. And so you do get the cost into a place that you could actually justify the return on investment. So you guys are participating in a lot of markets. You're into server virtualization, and you're into desktop virtualization. But they're not one and the same, are they? No, they aren't. Talk about the differences. So they're very different IO loads. And this is the challenge of virtualization in general, is you don't really know the IO load. You don't know what application is going to be running on a server. So on the server side of it, it is this unpredictable. It might be an application that requires streaming data at one moment and another application that's completely random IO in another moment. And so you have to be able to handle different types of IO and, again, deliver the performance for it on the server side. On the virtual desktop side, it's different. One of the real anomalies there is you see a tremendous number of writes. Storage in general is tuned for lots and lots of reads. And in virtual desktops, you have to be able to handle the writes as well. The thing we've been talking about, George, at last two days, day and a half, really has been around, obviously, desktop virtualization and VDI. I've been in the center of conversations. And it's all great. It's all sexy. It's all sizzly. People love to talk about the iPad, the tech jewelry that CIOs get, and they go to their IT teams. They make this work, right? But it's not that easy. I mean, IT has always been like a sausage factory, making it has been painful, but it had to work. But now, no one's talking really much here at Citrix other than you guys and a few others about the storage equation. Dave and I were commenting yesterday that everything comes back down to storage. And the nature of these apps, which is undeniable, the preferred user experience will be dynamic and require low latency data. What do people need to know about the storage equation relative to this new phenomenon? Because it's undeniably a trend. The app movement, an iPad, got it done. Windows, they have their own challenges. But still, at the end of the day, you're talking about a new kind of network behavior, a new kind of app behavior. Can you share some insights for the folks out there about what that means? So what it means from a storage perspective, because you're right, at the end of the day, data has to live somewhere. And without data, why are we doing all of this? And so it all comes down to the storage. And you're dealing with IT organizations. Their preference is it's your problem, not mine. It's your iPad. Go ahead and take care of it, just as the PCs were sort of your problem. And if you have a laptop, don't even talk to me. We'll give you a file share in some place. You mean the idea of bringing your own to work? Well, even if it was company supplied, the idea was they'd fix it if it broke, but recover your data if it was corrupted. No thanks. We gave you a file share. I hope you used it. But now, what's happening, of course, we have everything from legal requirements to just organizations living on that data. And when people leave, it's what, you walked away with the corporate data. What's the old IT expression? You give the end user a fork, they'll stick it in their eye. And then you're just... So it's like, it's about simplicity, right? But that's kind of the end user. They had the PC, those were the days. People call in, we hear it on the services angle. And number one thing is like, I lost my password, reset it. So users, they'll poke themselves in the eye if they give them a fork. So the new environment, talk about the dynamics of how that's changing for me, because everyone's talking about end user experience at the edge, correct, optimization. What has to happen at the inner workings around storage? Right, and around storage now that you say, okay, I do have to pull all that data in. I have to control, I have to manage that data. It means we're dealing with a new type of information that we have to track, we have to manage. It's not the traditional, you know, finance database. So we're dealing with a lot of data coming in that tends to be just random objects, I guess is the easiest way to describe it. So I need to have an expandable storage solution that's just gonna grow as the need does and is easy to manage. So in this area, the storage challenge is really growth and being able to do that as the demand picks up because what you find, iPad's a great example. We didn't have iPads not too long ago. All of a sudden, everybody I see has an iPad and so you start looking at the amount of data being generated. It's growing exponentially. So to say that I have, you know, I have 50 terabytes allocated to my end users, well, next week, next year, it might have to be 100 terabytes, right? We talked to one customer that had 300% growth in 18 months and so it's really the issue there is expandability. Yeah, Dave and I always talk about the impact of CIOs and, you know, Dave actually runs Wikibon which is the premier research firm, research analyst and they also do a lot of market intelligence and the CIO kind of consultancy. The number one thing on CIOs radars these days is a mandate for growth fast. So they kind of come out of this dark period of cut, cut, cut, do more with less to, okay, run. Run faster, do more investments. There's a mandate for investments. So, you know, the theme here is the PC era is dead. I don't know why they call it desktop virtualization. They should call it something else because they're basically saying the desktop's dead but the PC era is moving to the cloud era and so in the old days it was a scale up mindset, right? So, okay, scale up, throw gear at it, throw some hardware, throw some cash, build it, manage it, expensive. Now we're in a scale out mode. So you guys are very active in that concept of scaling out. Can you just tell us kind of one, what does that mean and what are the benefits of doing that? Sure, what it means is I don't have to buy a big piece of hardware and then fill it up and then have to buy another big piece because that incremental, you know, the next 10 terabytes that I need should cost me the same as the last 10 terabytes that I needed. So you want predictable growth. With scale out, you get that predictable growth. You add another, in our case, it's another storage system, another one of these bricks. Each time you fill up, you get to the next brick, you get to the next brick. So I can predict, from an IT perspective, what is gonna cost me on the next one? With us, one of the keys is not only can I predict what it's gonna cost, but I can tell you, I'm gonna have the same performance today as I have tomorrow. So that's another issue in the scale up versus scale out. In the scale up, as you put more and more load into that large box refrigerator, if you wanna call it, my performance is gonna change because I'm constrained by the controller at the top. In our scale out model, every brick has its own controller in it, has its own IO path. So I know that the performance I've gotten with two, three, whatever, when I put the fourth, I'm gonna have exactly the same performance. So it's that level of predictability that's so important. George, VDI and desktop virtualization have been relatively narrow, confined to certain industries and use cases, and there was a lot of question about the TCO as far as the mainstream goes. And so it was interesting to see you guys really go after the whole VDI storage. Of course, you did that post iPad, but I think it wasn't sinking in at the time. I think at a lot of the time people said, okay, that's kind of a nice niche, but it feels like desktop virtualization is going from this nice niche to now something that is very top of mind on every IT organization's list. What are you seeing, first of all, do you do specific, Xiotek specific VDI solutions, number one and number two, have you started to see a more mainstream uptake in those solutions? Let me start. The specific solutions are actually very important. Early on we talked to partners about VDI and why our storage is so great for VDI. And they got that, but they come back and say, yep, but you're only a piece of the solution. If I need to roll out a VDI solution, what are all the pieces, what are all the components? That's actually why we put together the VDI coalition. And so it's a group of companies who all have a piece of the puzzle there. We're working with Panologic, for example, with DataCore, with Citrix here, with Verstow. Glasshouse on there? Yeah, Glasshouse on there too. We had Richard Scanlon. Exactly. He was fantastic, he was great. So Glasshouse has actually put out a reference architecture for virtual desktops. At the show we announced- And he said, he commented, his business is booming because of it. I mean, he said, he's modularized his consulting practice, and this is a key part of that. I mean, there's consultancies out there, they need this, right? Absolutely need this, because it isn't an out-of-the-box. As a matter of fact, that's what we announced was pretty close to an out-of-the-box. Here at the show, a solution that you can buy with us, Panologic and Citrix as a complete package solution for people who want to start. One of the interesting things is, nobody deploys a 4,000-seat VDI environment. They deploy 50 seats or 100 seats, and then they go to 500 seats and then to 1,000. So it's a growing environment, back to the scale- You guys are strategic to Citrix. I mean, we talked with Simon Crosby on the queue on the first night on the kickoff for the queue here and on siliconangle.com. He's on stage right now talking about a keynote. He brought Zynga on board. You know, we asked him about storage, and you know, Simon's great. He loves to talk, oh yeah, yeah, the VM, wear this. But when we ask the hard questions about storage, he kind of doesn't have the answer. So you guys are the strategic linchpin there. I mean, because they don't really get in the weeds in the storage, they have all some cool tech at the top of the stack. But when they start to move down, this is key and their private cloud offering is kind of weak. So they got the net scaler, nice front door, back door. You guys are sitting there pretty with that little element right in the middle. I mean, is that the right read on that? And are you guys excited about that? We're actually very excited. I tell you, I kidded the folks we work with at Citrix that it was actually easier to work with VM wear than you, and that's quite an insult sometimes. You can suddenly get a backpack and then you get all kinds of concessions. Yeah. But what it was is it took them a while to realize what they needed from storage and to get some storage-minded folks really internally involved. And now that they have, the relationship has flourished. And they're open. Absolutely. They're open source based on the Zen side and the Citrix side, they know about partnerships. Absolutely. So we are working much better with them very closely. We just integrated to their logical volume manager as a way that we can add storage into the, with the Zen server, for example, without having to put controllers in the middle. So we can now basically build loans across these bricks as they expand out. So we're working a lot closer as to how we can leverage storage into the overall solution. Simplify, no controllers. I like it. Yes. So there's a lot of talk in the Twitter stream this morning about Flash. Simon Crosby basically said Flash changes everything. It completely changes the curve that the traditional storage curve is basically irrelevant now. Sort of what he was implying. You guys have a lot of core spinning disk DNA with Richie Larry and Steve Secola. You guys know a lot about that. What's your angle on Flash and its role in the whole desktop virtualization discussion? So actually we announced a product and shipment is next week for what we call our hybrid ice. And that is a mix of Flash and spinning disk. And the key there, speaking of Richie, some great technology he wrote as to how we utilize Flash with the disk. And we call it continuous adaptive data placement. Another one that just rolls off your time. CADP a little easier. But if you look at most companies, what they've done is they've taken Flash and they put it in front of their disk and they're using it as a cache. Relatively easy to do. Fits in with the current infrastructure, architectures, all of that. Problem is, you know, cache is sort of the, you know, think of it as a warehouse and as information comes in the front, you put it in the cache and then, you know, when it fills up, you start taking things out, right? That doesn't take advantage of Flash the way it really needs to be. Flash, matter of fact, think about it. What are the two issues with Flash? Well, it writes to it or slow and it does wear out the more you write to it. Well, what we do is we actually write to the spinning disk and then we promote into Flash based on heat maps and use patterns of the data. So what we're doing is putting the data that gets used the most, gets read the most in particular, into Flash. And so when you get one of our hybrids, it'll start real nicely at, you know, 5,000 IOPS. Within a few hours, you're up to 20 or 30,000 IOPS. Within eight to 12 hours, you're up to 60,000 IOPS that you get out of this 3U 14.4 terabyte unit. And so, thinking earlier, as I said, what's one of the problems with virtual desktops? It's IOPS. Well, now, you know, with 14.4 terabytes in a 3U, I'm getting 60,000 IOPS. So Flash is really big for us. We're actually really excited. Last week at TechEd, that hybrid ice product, won Breakthrough Product of the Year, and the Audience Choice Awards. It's a great, you know, sir. What event was that? That was TechEd, Microsoft TechEd. So, really good validation. We had Alan Atkinson on theCUBE, he was in town in Palo Alto for an event, and Steve Skola came by as well. He was talking about ice, obviously, you know, he was so excited, and he said, you know, orders are great, sales are doing fantastic, business is good. But then when I started digging in with Steve a little bit about it, it's interesting, you know, a lot of folks don't know that the Xyotec team, obviously with the new leadership with Alan on board, they have a DNA in this area. I mean, you know, out in Massachusetts, you got a lot of DEC, Digital Equipment Corporation guys. Give some insight to the folks out there who might not even be old enough to understand what Digital Equipment Corporation is, because, you know, they had a lot of this core virtual machine stuff back in the mini-computer days, a pioneer that strung on networks. Right. That's a redo now. Storage controllers. Well, conversion networking is just a, you know, smaller version of what was going on back in the late 80s. It's so true. I joke to people that if you look at that sort of the Vax world, you look at the mainframes, we still haven't invented anything new based on what was there in all these years. Data deduplication. Yeah, that's it. There, we got one. I got one. Data dedupe. That's it. Right. Everything else we've seen before. And it was probably more needed back then than it is today. But talk about what ICE is. I mean, because you got obviously that DNA from the background of some real, you know, networking guys, computing guys. They know computer science. And what's new now about ICE and how does that translate to the new model? So if you think about it, what Steve and the team were trying to do and your rights, this team has been together since, you know, the early digital days, the first rate systems that were there, the early storage works, you name it. And what they've seen are those trends and, you know, how describes if you think about, you know, the original platters that were out there and then how we got to SCSI and what that meant. And what the team looked at is what does it make sense to put even closer to the drives? So that's the difference between sort of the controller that we talk about that does snapshots and then provisioning and all of that versus ICE where you put the caching algorithms right there with the drives. You put the raid algorithms right there with the drives, not up on a controller in the server. And this data placement, the CADP, that's all right there with the drives where it makes a lot of sense and you can leverage it. Whereas the other functionality, which uses a lot of CPU and, you know, gives you different functionality, that now continues to live away from the disk drive. So we've gone to a new generation of a drive in many ways is what ICE is. Happens to be a collection of drives being that new generation of a drive. But because of the collection, the other thing we get is reliability. You know, two orders of magnitude. Higher reliability. Let's talk about this new generation of drives because when we talk with folks at EMC, Ropat Gelsing in particular, obviously he's been at Intel and he's on Maverick and he's been on theCUBE many times, half a dozen times. We asked him questions about security and Dave asked him, is it we in a security redo? And all this about discussion about a redo. Things are re-architecting. So talk about what that means next generation drive. Is it just a different kind of cobble together virtual drive? How should people understand this new? Virtual drive is not a bad way to think about it, right? If you think about the data pack that we call it, this sealed element, they're in the spinning disk version, there are 20 drives in there, okay? But you treat it as one and you never replace that one drive because of all of the technology in there to deal with the normal problems that drives have, right? Drives fail. You can look at percentages anywhere from two to 4% of all drives fail every year. Most of those failures actually aren't, right? Most of them can be fixed by putting it on a bench. And our goal is you fix them without putting it on the bench, it just happens. So that's what gets the reliability up there treating it as one. But it also is what gets the performance up there because you spread the data across all of the drives. So you're spreading data across all the spindles for each one and then you can read them off of that. So you're taking advantage of all of these spindles spinning at one time to deliver the data that you need. So it's that combination. And what is Citrix's interest in? You guys had an announcement, we'll talk about that in a second. You guys announced something yesterday with Citrix. And Citrix obviously is running fast. You know, they got the old school Citrix and the Zen team coming together and Netscaler. Great formula, possibly a triple threat. Yes. You know, apps, you know, personal and private cloud. Are you guys their store? They come to you and say, hey, we need some help. You get the phone call from Simon Crosby. How does that work? Just take us through that. So my phone wasn't ringing off the book, I wish it was. It is from, you know, other areas and customers. So it's great, but with Citrix, it's really been working with the organization because it is a changing organization. So, you know, we're working really more closely with the Zen desktop group than any group right now. And there's some great people there who really understand storage, who understand, you know, sort of a Veritas background and other storage companies. And they're the ones who understand what it means to work with a storage vendor. And that's where we started to have engaged. And that engagement is really leading to a lot of positives as to how we are working together and how we're doing it forward. And when they grow, they're going to stumble upon the storage problem. I mean, everyone gets there. Of course. Yeah, the day. Exactly. Put that back on the drawing board. Cool. And what are you announcing here today? So, a couple of things we're announcing. Oh, yes, I think. Yeah. So we did our Citrix Ready validation. We've had it for Zen server. Zen desktop was added, so we announced that. We announced this first solution from the VDI coalition, the Citrix, Pano, Xiotek, sort of VDI in a box ready to go and working with channel partners to get that into the marketplace. And then we introduced our MetaLun, which is, sits on the Zen server, uses their logical volume manager. That's the part that lets us take a controller out of the middle. And, you know, if I look at the long-term, you know, you talk about what Citrix needs to understand. To me, what Citrix needs to understand is that more and more of the storage functionality is going up into the server. So, if you think about it, the hypervisor as the operating system, is, it's going to do snapshots. It's going to do replication. It's going to do thin provisioning. All of that's moving up into the server, which for us is a great trend. You see it, you know, from really all of the, what I'll call operating system vendors, which includes Citrix, it includes VMware, it includes Microsoft, et cetera. Even database vendors, Oracle. Oracle's another great example of it. Function move up the stack. Absolutely. And so, as that happens, all of a sudden, this concept that I told you, which is we put what's really important close to the disk. It goes close to the disk. Everything else goes in the server. The middle disappears. And that is simplification. It is ease of management. And just a layer, you increase performance, too. I mean, significantly. There's a bottleneck that goes away. All right, good. We're here with George Simon, COO of Xyotec. George, I know you had another meeting, so we really appreciate you coming by and squeezing us into your schedule. Thank you very much. Oh, my pleasure. I appreciate being given the opportunity to talk to you. Love to, you know, anytime. You're in town, I'm in town. You guys are CUBE alumni. That's the new year of three now from Xyotec Day. It's been fantastic, so thanks for coming on. My pleasure. All right, good luck with everything. Thanks a lot. Okay.