 to completely change the game because it has several orders of magnitude more performance than disk drive and orders of magnitude more capacity than disk. Yesterday we announced our first refresh of our product line. We introduced a product four years ago that took us from 5x nanometer flash through 4x nanometer flash and 3x nanometer flash so three generations of flash lithography. Over that time frame this exact same hardware platform doubled in performance and maintained endurance even with the degrading flash. Now we're moving to the next generation platform 2x nanometer maintaining and improving endurance reliability but also more than doubling the performance level so this this matters because you can monetize the speed at which you can convert data to knowledge. Now VSL not withstanding VSL is you know how I feel about VSL it's a secret sauce that nobody else has I've been on record of saying that but EMC yesterday Pat Gelsinger announced that its PCIe flash card was in beta announced the beta availability to customers now that I I want to understand I've always asking I how big is Fusion IO's lead you know there's a lead Pat Gelsinger conceded that lead actually on the cue to his credit he was open about it. If you compare what they're shipping today from a technology standpoint again VSL not withstanding because that's not part of the equation to you know when you shipped your first beta first of all how long ago was that and what's the technology difference because obviously technology has progressed in those two or three or four years. So we've been in the market now for four years and OEM our product through HP IBM Dell which means it's been certified and hardened within each of those server environments and the support that VSL has been ported through Windows and Linux and VMware's ESX even Solaris were announcing with this last release OS 10 support for guys wanting to use it in a workstation space kind of a tangent market but fun and also support for HP UX on the itanium. So you know with this breadth because really when operating systems were first being built you had disks and memory we're introducing something that's not strictly disk not strictly memory and that means a very tight integration at the operating environment level. So we view this more as a question of the operating environment at the server and really what's going on here is storage has been serving two masters. One is the retention of data and the management of data through time so it doesn't get lost and the other one is the delivery of data in the processors so that processors don't sit idle. That data delivery versus data retention are two completely different demand drivers. With Flash you now have a way to completely bifurcate those demands and it has real implications for folks who have monetized the cross-coupling of those by charging on a per gig basis for performance. So that decoupling is what happens when you take the performance out of the storage array and deliver it in a much more effective fashion from the server because customers are not going to pay twice to solve the data delivery problem the performance problem that's why this is particularly you know a challenge for folks who've been making most of their business from the performance piece. So we think it's the bottom line behind all this is a huge market and having more players in and validate the market space actually is very good for us because there's a lot there's a lot of untapped market to grow into. So we had Pat Gelsinger up on the monitor here Mark Hopkins can pulled up the Pat Gelsinger clip and the blog post or one of those. So he did say as Dave pointed out that EMC's playing catch up the Fusion I.O. So obviously as a startup that now has gone public and got big EMC's big you guys are growing really really fast now public. How did that how did that go over well in the company? One you guys love that but you've got Pat Gelsinger putting his guns on you so like Pat Gelsinger is a fierce competitor he comes from the Andy Grove mindset you guys are funded and relatively compete but he's you know it's EMC they're huge. I think the statement is if you're not in the tank you're in front of the tank. I think that came from the EMC guys no in reality. Pat said if you're not out in front of the next wave you're driftwood. But no but he's a fierce competitor you guys have to compete you're now public so everything's out on the open you got other people nipping at your heels trying to be relevant like violin but EMC's obviously got project lightning how are you going to compete with all that? They did the same thing with Data Domain for the longest time they said oh that's a niche right it's a niche a niche we can do that better you know 2.1 billion later. The fact of the matter is this is our bread and butter selling this value proposition supporting this infrastructure you know EMC is definitely a fierce competitor but you know if the customer pays us a dollar to solve what would have been a $10 problem to solve in the storage array can EMC afford to undercut us even you know it would cannibalize their business more than us so the more noise they make in the market the more awareness that drives that there is a better more efficient way to solve the performance problem than paying the premiums within the storage array to try to get performance. So you're not necessarily in front of the tank you're on the side you're kind of growing in parallel to EMC because if they're going to cannibalize their business what that means is they really can't go after you so to speak right now they can they can if we don't get big enough fast enough they could go and lose money at specific accounts right yeah but that's you know we're kind of the Pandora's box that gets opened because we are taking a direct route to market because we have the reputation of making our customers successful with enterprise customer support it's one of the things that really differentiates Fusion.io and presents kind of a an enigma to the market is that you know people look at our products and say oh you're selling a component first it's the software that makes it into a full systems and solution but then more importantly it's the fact that we are out articulating the ROI helping customers change their infrastructure and use it and acting as a an enterprise systems company because this is an opportunity to solve the data distribution problem the same way client server computing evolved away from mainframes and led to you know the advent of companies like Cisco solving that problem. Well Oracle if they had their way they'd have a mainframe that's what Larry was basically presenting up there in his keynote but let's talk about Java 1 because right now it's Oracle Open Worlds here obviously Oracle but Java 1 is going on got a developer community MySQL they own NoSQL is the hottest trend in the unstructured market so there's this there's a big gap between unstructured and structured data people are trying to make sense of it do you throw more Fusion.io at the problem makes it faster unstructured is a new with new market how do you how are you playing in that with your products and what is your future outlook look like with that marketplace so that one could view it as kind of like three three separate things one is structured versus unstructured the other ones in memory you know versus versus on disk and another one is acid transactional compliant versus loose loose consistency models the advent of very high performance density flash address through the storage API's makes it possible to take legacy applications written to use storage and make them look as if they were entirely in memory without having to recode them because it is non-volatile memory so your persistence stays the same and you don't have to go and build it to distribute it across the whole bunch of machines so it lets you take legacy storage applications and immediately move them into a world where you can manage and manipulate very large quantities of data from a structured versus unstructured the the very high random access rates the ability to go and collect pieces you know makes a big difference versus having to sequentialize it and your data layouts on the media no longer matter as much and lastly with the consistency model the atomic right primitive that we just introduced helps close the gap so that you don't have to go through and double right and do other you know complicated mechanisms in the software to be able to guarantee consistency on the medium so obviously you guys are an exciting space performance-wise you guys are killing it you're doing great in the big accounts not under threat directly right now by EMC's the big guys but the world's changing you know we we see Facebook we see Apple we see Google doing all these cool things the new architecture of new clusters you mentioned client server we see in now unstructured I interviewed was at the storage networking world and he gets excited by new technology so you know Apple announcing today the iPhone 4s and iCloud and iOS 5 all this is going on they're now going to bring iTunes into the cloud Facebook obviously is a massive platform you guys succeed in these environments really really well what is the design criteria and can you talk specifically about why you guys are excelling in environments like Facebook I mean I see you know I wrote a post when you guys were in a quiet period around why you guys are successful outside of Facebook and it that more people going to look like Facebook in the future that more people going to move to a bigger cloud environment so Facebook Apple iCloud Google all these environments you do well in so why so can you share the insight it really boils down to a need and an awareness the the folks who are trying to accomplish a scale that hasn't been attainable before are looking for ways to solve that challenge and so there's a there's a you know somebody has to recognize there's a problem before they go and look for a solution and we have over 1500 customers most of them in just normal run of the mill enterprise running everything from payroll systems to you know doing movie editing it's it's very horizontal the solution because it just has to do with how do you feed data most optimally and into processing systems but customers don't come to look for us until they recognize that there's a need so that that can come because you're trying to get a performance level that that you aren't able to get that's an immediate one or it can come because you're not able to fit within the cost structure you can't just go and deploy another VMAX right so that the the the primary driver is awareness that there is a better way that servers are inherently capable of 10 times the workload they generally get today because they're starved for data in a timely fashion people don't recognize that fact because it's grown gradually over time to where there's just so scale okay so I buy that scale but the scale for these environments let's take Facebook for example Facebook is not your typical Oracle account me they might have Oracle that out there to run paid to make the paychecks get printed out properly in a workflow environment but you know they use my sequel yeah they use my sequel they don't use Oracle exactly my point is they don't use Oracle okay my sequel well not the Oracle in the sense that Larry wants well Larry Larry was very clear he wants to sell his own IP and Oracle's model is one vendor lock-in and own it right so you're successful in environments where there's diversity of architecture from commodity hardware to Hadoop and other environments that's an architectural change so talk about some of those factors around the scale because these are like the Facebook is scaling every day so why there well you know it is a powerful implement for solving these problem that is available to be integrated in a number of ways so I think there's a certain amount of flexibility there are places where we are used as a substitute for memory and actually under applications that were originally written to hold everything in memory like the caching layers of these web properties and then there are places where we're used as a as a faster disk it's that flexibility the fact that it is a very horizontal solution that makes it adaptable into these different environments David I wonder if we could shift gears a little bit and talk about some of the business decisions that you've made you and your board is as a team John and I were talking yesterday about the merits and lack thereof of going IPL shows to go the IPO route I mean you've seen Facebook and the their ability to raise money the valuations of private companies why did you choose to go public you know that there are lots of reasons to want to be a public company and for some companies those matter more than for others from from our point of view we looked at it because there are trade-offs there are costs to doing it for one you can't talk freely about things but the I think the main driver behind this for us anyway is as an enterprise solutions company selling directly to end users and knowing that it is awareness that there is a better way to architect information systems having being a public company where people can recognize your growth and realize that it's a real concern and not not some niche product of course that has the cited side effect of drawing attention of folks like like EMC to recognize that this isn't just you know a small thing but has has huge potential based on the growth rates so so that's our that was our primary thinking is to accelerate the business adoption because this is a it's a market capture because if there are like I said there's a lot of customers who simply don't know that their server systems are capable of doing much higher workloads than they get today and and I guess you kind of just answered it but but but same question around why NYSE versus NASDAQ you know that's that's a good one it you can tell us you can tell us that took you to better golf out of jail there is some brand value there yeah I mean there back in the day NASDAQ was cool for all the tech companies you know nowadays stability more stability to the yeah I was taking a tour at the New York Stock Exchange literally a month to the day after the flash crash as a slander on you know our dear products flash it was the it was the the big blip where the market dropped so fast it was very interesting you know to see how the systems that were in place at the New York Stock Exchange were ultimately what the SEC required of the industry to help solve the problem they already had it in place I also wanted to talk about Utah right I mean you know plenty of companies have had made it huge outside of Silicon Valley but Silicon Valley is the epicenter of all this innovation John as you know how did Utah come about and is that an advantage or a disadvantage for you why well there's an interesting story there I went to school in Utah and coming out of school I and two co-workers who are also in college at the same time went to work for Oracle started a satellite R&D office in Salt Lake City as part of the the effort to build a web browser so this was months before the browser war broke out between Netscape and Microsoft and Larry Ellison had aspirations to go into the browser this is the early 90s early 90s you got word perfect was out there too right at that just you know a few years ago that you had no villain word perfect who were two of the top four largest software companies in the world there is a lot of talent there and it's very centrally located you know not very far from Seattle not very far from the Bay Area here Colorado big storage hub great lifestyle business great lifestyle out there you know one hop down to Texas or partners down there so you know it's very central in that sense so I've worked my whole career from Salt Lake generally working for a Bay Area company became this satellite R&D office became the one of the seed engineering groups that founded NCI network computer Inc the thin client kick which is you know old things are all new again now we have VDI going that space so so today at Oracle it's kind of a slow day because of the Apple news but yesterday and the keynote big data is the center stage analytics caching a lot of people talking about big data caching are the two hottest themes okay caching you to mention a little bit about that you guys have a play a big role in that architecture big data analytics what's your view on analytics I see real time requires fast data it's on mobile devices that plays well to your strengths yep what's your view of the current state of big data and analytics and how do you see that evolving what's your vision and what's usually I owes vision for big data you know there there are contending visions for where that market goes you know it mainly revolves around and memory you know versus using storage systems and I think you know the the traditional ways of delivering data from disk drives it's known to be broken so most big data assumes you have to put everything in memory which means a big scale out model when you introduce flash into the market you now have the ability to use a more traditional scaled up model as opposed to scaled out in RAM so it has some real implications for how those software applications are authored to get their rapid supply of large quantities of data the interesting thing is that with IO memory flash attached like a memory device using memory controller and memory management techniques the the the cost per unit of bandwidth from flash is actually less than that of disk so most you know big data stuff where you're using disk drives you have to code it to use purely sequential access and it's all about bandwidth well even for just pure bandwidth flash when integrated correctly is now a more economic way to deliver to deliver the data so how about open world talk a little bit about the show what's going on here for you guys maybe compare it to last time we saw you was VM world it's amazing we kind of view the market through two separate two separate categories things that are inherently data intensive and those that aren't split it just like that the ones that are really data intensive are the ones that generally take multiple servers and and scale out or or you have to really scale up your storage systems and make them really beefy the stuff that's not data intensive inherently is exactly the stuff that you virtualize so that you can pack a lot of them onto a single machine right so the VM world kind of represented the efficiency where performance leads to consolidation and improved efficiency and here it's where performance leads to higher peak capabilities right so one is on the capability side of the market the other one is on the efficiency side and consolidation excellent all right David Flynn you know you always a great guest appreciate you taking time out of your busy schedule to come here and come back anytime we love having you on congratulations on the success so far it's like I say do it a hundred more quarters in a row and we'll be good congratulations to have you on the show David Flynn CEO