 for siliconangle.com, and we are theCUBE. This is our flagship program. We go out to the events, extract a signal from the noise. We go where the action is, and the action is at Oracle Open World in San Francisco. The center of all the action. Not so much innovation, because it's Oracle, but a lot of innovation around Oracle. And I'm excited to be here. We're here all day, Monday, Tuesday, Wednesday, day to day coverage. Again, I'm John Furrier, the founder of siliconangle.com and I'm joined by my co-host. I'm Dave Vellante of Wikibon.org, and this is theCUBE, and we'd like to welcome in Jack O'Brien, who is the Senior Director of Marketing at Violin Memory, a flash storage array company. Jack, welcome to theCUBE. Thanks for having me, Dave. Hi, John. Good to see you, and been the hot topic, John, of the show here, at least one of the two big ones, Cloud and Flash. So Dave, so we love Flash. Obviously we've been following the Flash from the very beginning, and then obviously Fusion, a private company that goes supernova, Violin comes out of Fusion, Don and the team, start a company. Then all of a sudden it's like the craziest trend, mainstream Oracle validating Flash in a big way. Jack, you guys are at Violin. You're setting the standard now in an area that is to be explored on the back end with Flash. Fusion's obviously in a different position. Both emerging companies, you guys are killing it from what we find and what we're reporting out in siliconangle. So first, my first question, well first, congratulations. So my question is, Oracle is embracing Flash. I'll see Larry Ellison on his keynote Sunday night, going all in on Flash, talking a big game, doing what Larry does best, taking credit for all the trends he created, and Flash is one of them, so not, which is the case. So you guys got to be pretty excited about that. So comment on one, his ringing endorsement of Flash, and how does that fit in to the real world right now as companies are clearly moving to Flash and in the architecture? Yeah, John, I thought you made a really interesting comment where you said Larry's endorsed Flash. This is probably the third or fourth time that he's done that. This is definitely one of his more emphatic times that he's ever done that. He came out and made an announcement around Oracle Exadata that basically said, the most important and new thing in this product is that it's got a lot of Flash in it, and you're going to run all of your databases in memory. And we've been saying the same thing for a number of years, right? My favorite quote from Larry's keynote was, well, disk drives are becoming passe. Now, becoming can mean a lot of things, but certainly it looks like we're ahead of the curve in terms of where the storage industry is going. And so that's what we're focused on. Enterprise storage arrays designed for the way that storage should be used, which is a shared network service that's highly available for your primary data. Well, tape has been going away for years. It's still the last kind of areas, kind of the media space. But yeah, disk is moving into backup. Flash is coming in prime time. But it's changing the architecture that people are used to using storage for. So there's an old way and new way. So describe for the folks out there, make sense of it in a very high level, old way and now the new way, in white flashes. If we think about disk drive technology, it's been around for decades, and that whole time it's getting pummeled by Moore's law, right? And so we need something to catch up and close that gap between there. And so what we're giving CIOs is, is the ability to close that one big gap that allows them to virtualize their entire data center. It allows them to put all their business critical apps doing things in real time with the reliability that they need. It allows them to take advantage of the various big data like initiatives to get the most out of their data, to translate that into information in real time. So Jack, I wonder if you could talk about violin's approach to solving that problem. So essentially what you just described, if I understand it, is taking advantage of a customer's existing processes and dropping in this new system, this new all flash system to speed things up. Is it conceptually that simple? Well, conceptually we don't want to disrupt their existing processes, their existing applications. And so the product we've designed does exactly that. It fits into your existing architecture. But where it's different is the approach we've taken with flash versus what others have taken with flash. We've designed a system literally from the chip up, you know, starting with our flash partner Tishiva, where we've designed our own flash cards and we were able to aggregate those in a very intelligent and dense way to ensure that you get the best performance and the best reliability out of a single system that is now a viable alternative to something from a top tier storage supplier like an EMC or an IBM or even an Oracle. So help us cut through this and squint through the sea of flash competitors. I mean, they're everywhere. But somehow you guys have merged. You got a cool name, so that strikes. But you also seem to be doing very well. How is it that you're being able to separate from the pack? Well, it's tricky to balance kind of these three, the three legs of the stool for enterprise storage requirements. You need to have the performance. Obviously that's the big advantage that flash brings to the table. But you also have to have the availability because people do not trust their highly valuable business critical data on something that may or may not be around or lose their data. And the third is you got to get the economics right. And so balancing all three of those things is very difficult. And our approach to architecting the system and the ground up solves those problems extremely well. So talk about that more. Maybe add some color to that. What do you mean by architecting the system from the ground up? Sounds good. What does it mean practically speaking? Well, I can contrast it to other approaches to solid state devices, right? You've got a MacBook in front of you which has a solid state disk drive. No spinning disk in there, baby. And that's one way, that's one approach to deploy flash. There are other ways, a common approach is to just take a disk drive form factor and put flash drives in there. It looks like a fast disk drive, but you're using it in somewhat of a legacy environment in terms of the controllers and the software and the things that are wrapped around it. If you architect from the chip to the card to the controllers to the system with very intelligent software and a lot of IP in the hardware and the software, you can design something that's highly optimized at the systems level. That's an argument that Larry would even make at the Exadata level. So talk about the software, talk about software a lot. Specifically, what does the software do? Well, at one level, it starts with the data protection and the parallelization of data. So we have our own unique proprietary rate algorithms that are patented that ensure that data is protected, that there's parity, that in the event of a failure, you're not going to lose any data. It complements the hardware portion of that high availability piece so that you can have full redundancy, hot swap ability, complete confidence that you're not going to have component failure to lead to loss of data. Okay, so essentially you guys are systems guys. We are. Hint to end, a lot of software innovation. Absolutely, it's not a matter of taking off the shelf components or other commodity components, cobbling together and wrapping them in some sort of envelope that looks like something you're used to. So to compare and contrast you guys to Fusion IO, I know you guys are completely different companies, but you guys kind of fall in that same genre. There's a little history there with Fusion and violin, but you guys are the upstarts. Obviously they went public. You guys have huge valuation on your last financing. So you're knocking on the door of public offering as we speculate, no official announcement yet. So no one will get you fired, but you know, if you want to con- Nobody's ever been fired from coming on theCUBE. Not directly. So comment on your IPO. And no, but seriously, compare and contrast Fusion. People see Fusion, they see, they understand what that is. They've been on theCUBE. Talk about violin. Well, in a lot of ways it is in Apple's and Orange's comparison, right? They're not building a primary storage array. They're building a card that fits into a server that accelerates the data that's local to that server. It's very different than enterprise class shared network storage. So that's really the fundamental difference that's going on. Now, is the underlying technology flash? Yes, absolutely. And flash can do a lot of things and there's a lot of different use cases where you might want to use flash. But while they're accelerating a single server is very different profiles for availability, for shareability, compared to what we're doing which is designed for shared network storage. So talk about the performances because we've done some kicking in the tires with some folks out there who have been playing with violin and in large production enterprise environments like SAP and others, where we're hearing feedback that it's literally smoking the doors off the benchmark. So, can you flesh that out and just share with us some use cases where you guys are crushing it on the benchmark side? Yeah, I mean a big part of our business revolves around making databases go faster. Batch and OLTP. If people can turn around a job that used to take 12 hours into under one hour, that's straight to the bottom line. Last week Cisco and Oracle released a benchmark, the latest TPC world record using violin on the back end. It's actually a trend we're seeing. Anytime somebody wants to create or produce a TPC benchmark, they look to violin to be the back end storage. Here's a great use case. Company that does all of the airline seat pricing. So this company called Pros, that's one of our customers. Their job is to take every flight, every class of service on that flight and every seat on that flight and optimally price it for an airline. If you can do that in a really fast period of time, you're going to make millions of extra dollars by optimally pricing your inventory. So people have been, I've talked to VCs going way back on the day, oh, Flash is like two years ago, oh Flash, just a caching layer. Now damn good caching layer when you look at what Flash can do. What else is there outside of that mystique? Because obviously now two years later, you're seeing startups, you're seeing a lot more broader, real adoption around Flash. Explain to the folks out there, what's beyond Flash? I mean what is Flash beyond just caching? Well I think that gets back to the point we were talking about earlier where Larry was saying that disk drives are becoming passe. And what that means is something's going to replace the disk drive as the primary storage layer. And that means you need to architect a solution that can stand up to the rigor that companies like EMC and IBM and HP and NetApp have been delivering for that long. And that's the approach that violin has taken and that's why you're seeing a lot of uptick in the enterprise as we've been talking about. So why not take that one step further? I got to have the conversation now. And write the applications to do things like atomic rights and totally change the performance profile. Well, there's a lot of things that we can do going down the road. I think that one of the things that's important is when you're delivering a new technology that can offer such a positive impact, you really want it to fit seamlessly within the existing infrastructure, especially when it comes to storage and people's data. I've called you guys, the data domain of Flash. Well, I mean, don't change the processes. It was really important for us, for example, to work with IBM SVC and EMC Vplex because those are tools and frameworks that enterprise companies have been using to measure data and they're very confident in them. We don't want to tell them, yeah, you can't use that or our new startup software is better than theirs. You know, that's not the road to success in protecting somebody's primary enterprise. Well, I think they can both be successful. I mean, there's a big market out there. Exactly, exactly. We've seen one of the, I guess up until now, about a year and a half, two years ago, one of the greatest wealth creations in storage history with 3PAR and Component and Data Domain. And do you think Flash will be as big? Well, we're talking about, what, a $30, $40 billion opportunity. And there's a huge, huge percentage of that primary storage that can benefit from running faster. And so I think that there's a big opportunity there. So from a spending standpoint, do you expect 10 years from now more to be spent on Flash than on spending disk? I don't know where the tipping point will be from when Flash to disk kind of passes. Well, I went out 10 years, so that's kind of a long time. So let me answer that question this way. What we're seeing is we don't necessarily come in and tell somebody who bought some storage system six months ago, you got to get rid of it. But as an enterprise is looking for a new project, maybe they have a huge virtualization initiative, maybe they've got a big data initiative, maybe they're looking to move some critical business apps to make another line of business successful. They're saying, okay, we're either refreshing or starting new, they're all looking at Flash. Right, I mean, you talked about the MacBook a while ago, I mean, just as a consumer, John, I'm never going back to spending disk, you know, relatively new. Well, I just know, Dave, there's no doubt in my mind that Flash is going to replace disk. It just absolutely has to happen. Certainly for active data, right? There's no question. No, no, I'm saying, I think Pure Storage came on theCUBE and said, you know, he didn't say, Storage is passé, he said it's a truck and replace it. So EMC could say all they want about, okay, you know, a small bit of Flash. This is the road that we're on. So the whole adage, a little bit of Flash goes a long way. Yeah, that's today. Your tweet was, oh, what about a lot of Flash? Yeah, so, you know, all the top companies that are need performance, whether it's Blacko search engine, number three search engine, or whoever, they're all putting Flash where it matters the most. It's expensive now from a unit cost, but in the aggregate scheme of things, when you can change your infrastructure OPEX and drive revenue, it's a small, different cost of ownership equation. In three to five years, we're going to see Flash dominated in the primary storage layer. Yeah, there's no doubt about it. And I think what's exciting to me is that, you know, we brought this up on theCUBE at VMworld, is the software developers and the DevOps folks, however you want to look at it, OpsDev, DevOps, and or software guys, when you have that much quote memory to work with, I'm not going to say RAM, but you know, it's not all state Flash, you have a lot of different dynamics going on as a developer. And I think that's going to be interesting to watch. I think this will be an absolute explosion for the DevOps community and the data center and then in the big data infrastructure, I mean, software, it's software led infrastructure. It is, it's a term we've been playing around and essentially, I wonder if you could comment on this, Jack, I mean essentially we're seeing, you know, the last spinning bottleneck, the mechanical, electrical mechanical bottleneck disappearing and we have talked about new architectures that need to come forth, where all active data will be on Flash multiple tiers, at the server level, at the cache level, the all Flash, even within the array, certainly hybrid arrays, but the dominant position, you know, being Flash, the primary storage, as John just predicted. Exactly. What I want to know is, can you give us some more color around the benchmarks? Because again, you know, I don't think people can truly get their arms around how much advantage there is. So you have any like anecdotal like data, like, you know, application, you know, an SAP application takes, you know, five minutes to run, now it's five seconds. Is there, I mean, we're looking for order of magnitude because we're seeing those kinds of, I mean, I made those numbers up, but just to compare the five minutes to five seconds, I mean, that's a notable, you know, fall out of your chair experience. Right. So do you have any data on that? Well, I mean, so, you know, without, you know, without talking about specific customers, we can see scenarios where we go in for a test with a customer and it's, okay, let's just see how the array can do if we hook it up to our existing application. And you know, without any tuning, we can double the speed of something. You can look at how Cisco decided that their performance benchmark that they were going to highlight, you know, a couple of days before Oracle Open World started, was run on violin on the back end. When we were at VMworld just three weeks ago in this very conference center, VMworld got on stage in their keynote and said, hey, we just put a single VM doing a million IOPS. Guess who's storage was on the back end for that one. So you can kind of see the trend here. If you want to get the best performance. Get a little smile on your basil smirk there. If you want to get the best performance you can, you're going to be putting a violent array in your solution. Well, I mean, I think you guys got a good business. So give us an update on your positioning one. Obviously, yeah, how you going to market and what's the plans for this year? What's the, how's the business outlook for you guys? Well, our positioning is pretty straightforward. We're primary enterprise storage and we're the only company that can really have a solid foundation on all three legs of performance and performance that's sustained as well as high availability and economics so that you can truly replace disk with flash in your primary data layer. You know, the example I used to use when I used to be in the sales business, so you used to go out to customer accounts and you know, call it the house on fire, the emergency room, the triage. You say the house is on fire. You go to the first room first that's burning up. You want to save that room. When you talk to customers for violin, what are the areas that I need you right here first? What, can you break us down, kind of sequence out the areas that you kind of hit first, you know, to save and or improve infrastructure? The first would be a business critical application that you can directly tie to the bottom line of the company. So say you are a financial services company and the faster you can analyze some data to make the smartest trade, the more money that goes to your bottom line. It's obviously a business critical application. Violent absolutely needs to be in the mix for that. Everybody's trying to go the next, you know, 50% of virtualizing their entire data center. The easy apps have been virtualized. Now the hard apps need to be virtualized. And to do that, the biggest roadblock is the storage IO bottleneck. And so violin is making a lot of strides in there. So it's the performance issue. Well, it's the point you made up before. It's kind of one of the biggest bottlenecks that you have to deal with. And we can open up that bottleneck in a big way. And application development heads don't want to virtualize those apps because they'll risk slowing them. Well, actually one of the interesting trends we're seeing is a lot more real large scale production VDI initiatives. And part of the reason is because flash stores can make those actually painless. I saw you guys just had a big win there, right? Yeah, we've had a couple. There's a university in the UK that decided to go all VDI for their student computer infrastructure. The largest air transportation services firm in the UK went all violent for their VDI infrastructure. So it's one of those things where now we can do things we couldn't do before. You know, and that's kind of the secret to why violins really being successful is we give enterprises the opportunity to do things that they couldn't do before because of technology roadblocks. And we're getting through there. What do you like at the show here? So describe to the folks the vibe here. What's your takeaway from Oracle Open World? Also, so the keynote you've seen, what's going on? I actually find, one of the things I find really interesting is that even though it's an Oracle show and Oracle is known as a software company, there's a lot of infrastructure and hardware discussion that goes on here. A lot of people that know what they're talking about, a lot of architects that look at the big picture. In fact, some of the toughest questions we get are the guys that know the database, know the application layer, know the middleware layer and know the hardware layer. And those guys ask questions that tie all those things together and those are the coolest people to have come by your booth and look at your product. Violin Memory Systems, great company. We're following them, they're right off 101. Great leadership team in Donbassil, the CEO. Great, great, aggressive company, but they know what they're talking about. Pioneering and Flash, Jack O'Brien, thanks for coming on theCUBE, appreciate your time. And congratulations on a very successful Oracle Open World in the sense that the leads is just going to start flowing in. Even Larry just endorsed your vision for Violin. Congratulations and thanks for coming on theCUBE. Thanks John, thanks Dave. Okay, we'll be right back with SiliconANGLE.tv, theCUBE right after the short break with our next guest. Thanks guys. Thanks.