 Okay, we're back live in Silicon Valley at Storage Network World. We're here with special guest Val Burka Vici. You got it. All right. All right. He is a, I'm going to say chief scientist, CTO, strategy, but he's a tech guru at NetApp. He was in marketing for a while, but they kicked you out, right? Yeah. He's in like a view. He's like, he's in research, he's sniffing all the trends. We're at the largest trade show conference in storage and or infrastructure called SNW, where a lot of action in the storage business is changing, a lot of new opportunities. This place is growing like crazy. It's an explosive market. We're seeing the likes of Facebook, Twitter, Apple, Zingo, all these new applications, new user experiences, all kind of being enabled today by data, by new infrastructure. And storage is an industry that's changing from disks and spindles to storage farms to actually being a core part of the engine of infrastructure. That's creating massive amounts of wealth, great user experiences and new technology. So Val, welcome to theCUBE. We're going to extract that knowledge from you and get it out there and share with everyone. Dave, Val, just quick. Big data is a hot trend. It's changing the storage business. We were talking before we came on live that virtualization, been there, done that with NetApp. You guys are doing it with VMware, Citrix, whatever VMs you can do. You can get your hands on. It's kind of old. But now there's a new tornado happening and that's around this big data, something that Dave and I have been tracking very heavily for over a year now. And now I want to dive into that with you. So first, tell us what's going on in the marketplace right now in this SNW, the South by Southwest of storage we were calling it earlier. What describe to us what you're seeing out here? A couple of really interesting and timely trends. I think probably the first thing I'm seeing is that there's a lot of theory about cloud at SNW and other similar events last year, but not a lot of adoption of real cloud-like deployments, whether they're private, public, hybrid, et cetera. This year for sure, I'm seeing real practitioners, real adoption of cloud technology under various categories and the names are becoming less and less relevant to me. It satisfy the characteristics of economy of scale of self-service is charged back, so it feels a lot more real to me. As I mentioned to you earlier, I see a real distinction between the kinds of applications that being used in the cloud that distinguish a traditional cloud. I'm calling it cloud one, it's a term Mark Beanieff and Bill Coleman and some other industry vision areas are using right now. Cloud two is really all about a new style of application. And you're seeing a lot more prevalent in public cloud services, services at scale particularly, I think of virtualization not along a machine boundary, but along the services boundary. So various elastic units of compute, various elastic units of networking, and certainly various elastic units of storage with different policies and profiles, that is really what cloud two is all about. And big data is a huge part of that. One of my strong beliefs though as a former application developer is that data by itself isn't very interesting. There's a lot of new code being written. New applications being written to drive value out of that new data to literally monetize it in many cases. And new code doesn't happen in a vacuum. And actually nowadays is being deployed on new platforms. Platform as a service is hot. And that combination I think of platform as a service maturing in 2011 is really the coming out here for platform as a service. Combined as a storage company with a lot of data, big data, is that hot confluence of trends that I'm seeing drive a lot of activity. And the startup community is responding. We're obviously here in Silicon Valley where they host us for I think the first time or the first time in a long time. There's a lot of innovative startups out here. A bunch of net X net app folks we talked to earlier. But this is really innovative right now. And can you give a peek for the folks out there what you're seeing in terms of the key innovations are from startups and toward kind of what net apps thinking about? Well, I think a lot of the key innovations interestingly enough are certainly coming out with how do we optimize virtual machine hypervisor based clouds? So whether it's VM specific storage solutions, certainly the maturity of a VM oriented data protection solutions. At net app we have this term, this engineering term we call VM granular data management. So it's being able to natively operate with a VM unit of granularity for everything you do. Whether it's provisioning, some capacity, whether it's tweaking it for performance, whether it's setting policies for availability, for recoverability, recovery points, recovery times and so forth. Using that as a core unit of granularity is really, really important as you have VM sprawl or just as you're starting to manage VMs at scale. So I'm seeing that certainly as a big trend at this show. And I'm hearing a lot of talk, certainly a lot less product delivery but a lot of talk around next generation application platforms and again that next generation of virtualization which is all about much more scalable platform based environments and big data type infrastructures. So I haven't seen a lot of for example Hadoop oriented announcements at SMW this year and I expect that to actually change quite a bit. Is that because of the vendor list that's here? Is it because of the people that are here are kind of, I don't want to say clueless on Hadoop but Hadoop has really come out of the open source community. Obviously as we know, yeah. It's a developer led initiative right in the middle. Yes, it's developer led. It's very new, quite frankly, particularly the storage people and I see actually some really clear distinctions, organizational distinctions that explain a lot of activity here. So this is very much an IT operator kind of show. It's you know, I have to operate IT, I have to purchase storage, it's optimized for operating applications that I'm used to. So whether it's custom applications, whether it's packaged applications, PeopleSoft and so forth, SAP Oracle, I operate that environment. I don't write those applications. So I need solutions that help me maximize the efficiency of that operation practice that I run. Hadoop is an example of various, you said Dave, an application driven trend, an application developer driven trend. And what's really interesting right now is I've seen a pendulum shift between the amount of business value and operating a packaged application is diminishing over time quite rapidly now. Whereas the amount of business value and monetizing, all the information we're collecting. And think about the sensor data alone, all the RFIDs we walk by, the barcoding, machine generated information, smart meters and so forth. There's an overwhelming amount of information right now. And you can either just drop it or you can actually keep it and extract it and mine it. Sometimes you mine it in real time because you know what you're looking for. But interestingly enough, if you're able to retain it at a low cost, then you're able to mine it over time and find value, latent value you never knew was there. And it's not sample sizes anymore, it's not estimates, it's all your data. And that's what's exciting here. It's everything. It's the real deal. It's entirely the fingerprint of business and or anything, right? And we sort of call it as, there's a lot of HPC techniques being used to do that from a technical perspective, but now instead of being relegated to scientific labs or a couple of specific genomics applications or whatever you're seeing. HPC going mainstream. And that basically is one of the ways I describe Padoop from a business context. And Dave, we talked with some of the entrepreneurs out there and one of them that we like is Marco Pacelli with ClickFox out of New York, not known in the Silicon Valley circles, but it's like a fifth startup. He talks specifically about that business value and that the data mining actually has, for the first time, clear business objectives. So the elusive, what's the ROI that's been this phantom or prefabricated metric is now real data and that you should have it available and that people who don't do it are going to be kind of either out of business or... And you mentioned Marco and he even made the statement. It's really not about the infrastructure. It's about the applications on top of that. There's no doubt about it. Even Dave Hitz admits, hey, we're the plumbers in the industry and but when you have cold, cool water going through the pipes, you're happy when you don't, you're really unhappy. But Val, you told me last night that you've been following the Hadoop movement for almost three years now. Yeah, since the very first Hadoop. That intrigued me because Hadoop is not on top of mind from mainstream for most storage companies. We were at Hadoop World last year and there weren't a lot of storage people. That was one of the very few. I know you had some people there. So what led you to sort of track Hadoop and what does your sort of trend spotter DNA tell you about that movement? It was really, really simple. I mean, at NetApp, if I don't remember it myself, my bosses remind me all the time, we're not a charity, we're a for-profit company. And we have an addressable market. It's very much targeted towards enterprises that have to operate their infrastructure as efficiently and responsibly as possible. So what I looked for, because there's lots of interesting technologies out there, what caught my eye with Hadoop going back three years was the fact that there was an event, Hadoop World. Was the fact that it wasn't in Silicon Valley, it was in New York. So they are already an implicit business mentality to the event. And most importantly, even in the earliest days, there were already some, what I consider, bricks and mortar mainstream enterprise use cases being highlighted even three years ago in the New York Times, basically reform and new transformation of their archive, digitization of their archives, was the one profile example. So that caught my eye and I thought, hey, you know what, I'll put that aside as something interesting and something worth watching. The year after, there was an explosion of enterprise use cases built on Hadoop. Mainstream bricks and mortar businesses, not just web 2.0 companies with better ad placement. That doesn't gain a lot of credibility in my business discussions inside NetApp because that's a niche, that's not a market. You know, fast forward to year three and you had 1,000 plus attendees, maybe 1,200. Yeah, it was mostly enterprise from what I can tell, not necessarily just web 2.0 guys. And all of a sudden, you know, the dollar signs are going off and everybody said it's watching. So it was really that bricks and mortar that not just West Coast web 2.0 association from the very earliest days that caught my eye and with every subsequent year of activity reinforced my notion that this is going to become infrastructure in business going forward. So you watched the early adopters, but for you, it was the fact that it went from early adopters to in practice because the proverbial chasm was crossed. Yeah, and a specific kind of early adopter. It wasn't, again, it wasn't just web 2.0 companies. It wasn't just Facebook or Twitter or Google or whatever that was using this technology. It was the fact that enterprises across multiple verticals, retail, finance, healthcare, energy, they were finding value in this technology, that's what caught my eye. So it's a different game, though, for a storage company. Like how do you, you know, your management can remind you with your for-profit company. So how do you profit from Hadoop where you don't have this sort of put it all in a box mentality? Everything's distributed, you're bringing the code to the data. What do you, I mean, it's early days, but can you make money at Hadoop? I think so. In fact, I'm doing more. I'm actually making investments specifically, and I'll give you one example that relates to a little bit in an open source context. When we first at NetApp started seeing a lot of customer traction with production databases over NFS, for example, over Ethernet, on Linux, we found was a lot of vendors had very mature, robust NFS server implementations. NetApp, of course, had one. Sun NFS server was solid, HP NFS, IBM NFS server solid. But to run a database from a host, it was a client portion of NFS that was being stressed. And quite frankly, the client side of NFS was embarrassing from both Sun, HP, IBM, and we had no influence to control that because we needed a stable client implementation to be able to support production databases. So what we did is we basically started aggressively investing in contributing to the Linux kernel. And we now are the lead maintainers of the NFS client and server portions into the Linux kernel by throwing money at the problem, by investing engineering resources. And we shamed, effectively, by having a better Linux implementation of the NFS client, Sun, HP, and IBM, to name a few, into upping their game and having a robust production quality NFS client so that we could have an end to end connection. And you were an accelerant to that process. Yeah, we had to break the chicken and egg cycle and we had to invest. The very same model can be applied now to any other open source project including distributions of Hadoop. Particularly since Hadoop is based on the BSD license in Apache. So any IP that you contribute to Hadoop, you don't necessarily have to share if you don't want to. If there's something that you think will differentiate you and has distinguished IP there and intellectual property in that context, you can retain it. And I think that is a very business friendly, commercially capitalist friendly model. And that's why I think that whole Hadoop ecosystem is already thriving and it's quite diverse. It'll continue to be so as companies are adding value and then they can monetize and retain some of that as a proprietary advantage as well as share others. Is it a war or is it collaboration? So at the big data conference, which we giga owned, had a one day event in New York City, we had a big data special that day and we were also running our clips. So we were kind of plugged into that. There was a lot of discussion around competition. So there was a series of announcements, data stacks, among others, competing against Cloudera. So I wrote a comment on Derek Harris's post on giga owned business saying, do we have to compete? Does it have to be a war? Because as you know, in early ecosystems like this, it's industry formation. So there's a lot of money to go around. It's not like the market's small. I mean, it's a big market. The Hadoop, he said, it's all the verticals, financial, government. You name them, they're all going crazy for Hadoop. So the question is, what does the industry formation look like? It's Apache based, it's BSD licensed. So there's investment, BC, big company investments, startups. What's it going to look like? I was going to prognosticate here obviously and just speculate, but I think it is going to follow a little bit of the Linux model, even though it's a different license there, but if anything it encourages more commercial adoption. But fundamentally, despite the politics and the open source movement, it's still very much meritocracy. So I really believe that companies and organizations that invest the most in contributing to whichever distribution of Hadoop ends up becoming the ultimate one. And maybe you can live with a top two or three, you don't have to have just one. That will ultimately be what most enterprises or most organizations, NGOs and so forth, ultimately go with. So I predict standard Darwinistic practices for any industry where the strong will emerge based on meritocracy, really the most contribution to the distributions. And there might be like Red Hats, say one or two dominant distributions. Fragmentation will only keep the industry lagging behind and I think people will get over that pretty quickly. And where do you play? So it's a more software potentially, right? And it's solving problems with the performance or management or data protection or whatever it is, movement. The problems don't go away. It's the same fundamental problems. It's just on a bigger scale right now. So scale is a difference between technology we sell today and technology we're selling into cloud two into Hadoop type environments and the ability to apply solutions for how to keep these mechanical spinning brown rust discs at scale operating for how to actually manage the data, for how to protect it, for how to make sure that you have availability of the data and resource efficiency and utilization is a really important thing as well. We're seeing everyone focused on scale today as they should. If you have a viral application, your number one concern is keeping up with the demand. But after a while you hit a plateau and then the joke within the company is the green visors show up. And I say, okay, now that you've hit your plateau and now that you've kept up with all the demand, there is, you notice there aren't a billion Facebook users yet, right? They're still at 500, 600 million only. But they've hit a plateau. All of a sudden efficiency matters. And that's where the IP we have and the portfolio of efficiency technologies will be applied at scale. I'm not saying we have all the solutions shipping today but those are the kinds of things we're working on. And those solutions at scale will solve the business problems and that were opportunities before. So you see the fundamental tenets of NetApp's efficiency approach. I mean, we heard, we hear a lot about it, right? I mean, it's the D-dupe and it's compression and it snaps, et cetera, et cetera. Those fundamental capabilities you see actually carrying over into the big data movement, maybe in different forms. Exactly. The qualifier is at scale. You can never forget scale in this world. And yeah, the interfaces are different, right? It's not an NFS interface anymore. It's Java, RPCs basically is what the Duke file system is. So the interfaces are different. The core value is the same. And you know, as long as you keep the focus on scale and you're able to be cost effective at scale, responsive at scale, supportable at scale, you have a viable solution, a detraction solution in this market. So you mentioned, essentially, you've got to pick your spots in terms of where you put your resource. How about database? What's your vision there? What are you seeing? I mean, you guys going to go out and buy a database company? I mean, you know. Unlikely so. Vertica, great fun. I don't know if you've had to chat with Jeffrey more recently. I had to privilege a while back actually at a previous, maybe it was even SMW show, or no, sorry, it was the fall show storage developer conference. You want to get him on the cube? No, not yet. We'll get him on. Get him on, because he's got a couple of great terms. He describes cloud one and cloud two loosely as systems of record is cloud one, which is very much a database environment so we're familiar with today. Column legacy, if you will. But you know, there's still a thriving market like at Oracle's results. And then there's systems of engagement, which is really all about interaction, collaboration, big data monetization and so forth of a lot of, you know, big data sample, big data sizes versus sample sizes. So Jeff has this wonderful term. We don't replace old systems. We pave over them, right? And look at the BIOS that we all thought about when we were running DOS. That was really, really important. No one, and BIOS is still there, but it's just a detail now buried so far underneath OS layers, driver layers, middleware layers, database layers that we just don't care about BIOS anymore. Same thing is happening to transactional databases. Transactions matter, you know, when we transfer, when we buy stock or transfer funds, we really want to know that that very simple from a discapacity perspective, you know, record, text, alphanumeric oriented transaction, it has to occur, it has to occur fast, it has to be atomic. But from the grand scheme of things today, in terms of, you know, all the new data types of video, the social media, the mobility of information, it doesn't, you know, it's not as relevant a portion of the IT spend anymore because it's a small, literally, it's a small amount of capacity. It's a relatively finite size system that can process that, typically a scale up system, whereas all the exciting new business innovation now, enabled by big data, enabled by video, is all around, you know, larger infrastructures that are, you know, scale out as opposed to scale out. You guys, obviously, are for profit and you serve as enterprises, but a lot of enterprises serve end users. Yes. So there's a big consumer applications out there from, you know, whether it's a web app or a financial services company, the consumerization trend is obviously here now and present and talked about for a decade. How do you look at the notion of personal cloud? I mean, we all have the idea that, at some day, we'll have a cloud of our own, our own hybrid cloud. Many of us do. And we kind of do it now. I got a little Facebook here, I got a little LinkedIn here, I got, you know, Twitter here, I got my hard disk on site, I got Roku channel, all this stuff's going on. And this ephemeral data trend is real. So like, that brings up things like, okay, that's an interesting concept, but what technology drives that virtualization, desktop virtualization, conversion networking? So all this is kind of coming together. So talk about what your angle is from NetApp and how you see personally the desktop virtualization layer and what's going on under the covers there. So I'll start with actually a business model answer to that question and I'll get to the technology. We realized early on we have repeated requests about two years ago, maybe even two and a half years ago from customers for NetApp to provide a storage service in the cloud. We thought long and hard about it, you know, could have been certainly a whole new department for us, a large incremental market opportunity. And we realized we don't have core competency in the services business, especially the online cloud services business. And ultimately we have this term that cloud will result in an aggregation of service providers, but you basically have, you know, people that become dominant as this industry matures and we just don't want to compete with ultimately we think are our best customers today and we hope will be, you know, even more of our best customers in the future. So our business model is very much around enabling service providers to be NetApp powered, but not offering NetApp, you know, native NetApp storage services. We'll offer branded NetApp branded storage services via Terramark, T systems, Rackspace, other major providers of ours, BT, you know, France Telecom internationally. So we have a NetApp powered program where our salespeople are compensated neutrally with the same accelerators and quarter retirement, whether they sell gear, kit, or whether they actually sell a storage service based on a NetApp powered service provider. So that was a major effort. You know anything about sales compensation or anything? That was a major 18 month effort. And again, the technology to enable that is in my mind, split along these cloud one, cloud two boundaries. So Jeffrey Moore term systems of record, it is very much about server virtualization. It's about desktop virtualization and that's evolving towards application delivery now, more granular virtual desktop virtualization and so forth. And these new applications now that are driving big data are again, a different stack, a different set of virtualization assumptions. And those are, you know, some of the areas of investments that we're working on right now. Like specifically, like what do you see there? What kind of changes in virtualization? Analytics is something interestingly enough, you know, it's the traditional data warehousing market is not a market we chase aggressively. Our design center, our core technology is optimized for response time and latency, not for bandwidth historically. Hence the ingenious acquisitions of a bandwidth optimized play. Teradata automatically becomes, you know, a major customer of ours and a major partner of ours, IBM's other storage product line aside from N-Series. So we're basically seeing these new style of applications that are not replacing the old ones, they're just complimenting them, maybe paving over them, certainly overshadowing them over time in terms of growth of the dollars spent. Those are the interesting new markets that we're getting into. And again, whether it's a hypervisor, virtual machine based optimization value that we have today and very much in that tornado. In fact, we've crossed the chasm we're into mainstream adoption of virtualization technologies beyond the low hanging fruit into the business critical and mission critical virtualization phases as Meritz describes as where we are right now. But at the same time, we're seeing this growth in custom application development and enterprise. Some of these applications being deployed as cloud platforms and internal platforms. And what excites me is the new stack that's supporting all these things. It's focused on scale, that's focused on technology such as Hadoop, focused on real time analytics databases. It's new. I'm bored easily, so I like new stuff. So talk a little bit more about the ingenio acquisition because you're implying that there's a big data angle there. Talk about that. First and foremost, so just to be official, it hasn't closed yet. The transaction we expect to close sometime in May. And we'll have a lot more to say obviously publicly once we do that. But what we have said so far in terms of justifying the acquisition was first and foremost video. Video was not a design center for a waffle and on tap, particularly high bandwidth video ingest. If you take a look at some of the applications right now, particularly in the classified side of government, you'll notice that they require anywhere from about 10 to 50 gigabytes per second of sustained throughput. Wow. Wow, exactly. That's a jaw-dropper. So it's not about just continuing to improve the engineering performance of our product line. That's massive. It's a fundamental step shift. And we know that from USP 2.0, move 10 megabits per second. 20 hours takes us 20 hours to move 600 meg, I mean gig to back up. So some of the predator drones out there to use as one example are collecting a lot of video information and that needs to be ingested and pre-processed in real time. And then certainly some retroactive processing as well. So that is a giant, giant market. It's a large base growing as exponentially as every other data type. But if it starts off a large base, and again, this is now video becoming more of a mainstream requirement as opposed to a niche, that motivated us primarily to look at the ingenial property and to make the acquisition. Secondarily, we also looked at the fact, as I said earlier, HPC is going mainstream. And Hadoop is just an example of that. It's shared nothing, massively parallel, distributed storage technology, but it's being applied to a lot of business problems as opposed to research academic problems. And it turns out the ingenial product line is again, as optimized for that in terms of high performance, low cost, dense storage, characteristics that we didn't optimize, you know, Waffle and ONTAP for. So it's completely incremental market opportunity, completely incremental workloads. And quite clearly, you'll never be seeing us position, you know, ingenious as a SAN alternative to ONTAP and Waffle for Oracle, for Exchange, for VMware. We're gaining so much share in those markets with our existing technology today that would be foolish for us to do that. Yeah, so, and John, you know, I like to pontificate from a financial standpoint. I mean, I was pretty vocal about EMC's decrease in core value last year. In VMware, in the market cap. We've been, and we've seen some changes. One of the things I've said about the ingenial acquisition is essentially NetApp picked it up for, I don't know, less than a dollar, you know, on a revenue basis, less than one X revenue. Okay, but NetApp's trading it, let's say, for maybe, yeah, four times revenue. So the day they do that acquisition, their investors get an immediate bump. I've said at least three to four times. If they're paying $400 million, you're talking about a billion dollar plus valuation hit, just for that acquisition. Assuming they can keep some of that revenue, but there's a lot of, I mean, George, and we asked George is that on theCUBE, Tom, the CEO of NetApp, you know, about innovation, about M&A. You know, you're going to grow organically innovation, but this is the deal, and LSI was, you guys recently got involved in that deal. These are bargains. I mean, you can't pass up a killer value, right? I mean, is that kind of the philosophy? I mean, of course you're going to do innovation. Well, it's an undervalued asset. It is a great Warren Buffett play. You got to get a good brand. I mean, maybe not a 75-year-old brand, but in our world- That's a good headline, though. It's almost a Warren Buffett approach to tech investing, which is rare in it. It's undervalued, it's a value play. IBM makes those same moves, too. I mean, IBM is known for making those kind of moves. So I'm presuming your assumptions are you're going to increase revenue. It's not going to decrease revenue, but even if you do, it's still from an M&A standpoint a good deal. So I'm on record of saying it's a smart, financial move, again, assuming you can manage it right and grow the revenue, it's a good dub. The markets you mentioned are big markets. I mean, if that truly materializes into the value you're mentioning, big data, cross-multiple verticals, given the fact that that was an area that you needed to have an HPC standpoint. Yeah, yeah, it should do well. You won't lose customers, you'll gain customers. We predict the full-motion video opportunity alone over the next couple of years justifies the acquisition by itself, regardless of the large OEM revenue that's already part of the acquisition, regardless of HPC going mainstream in Hadoop. Full-motion video alone justified the acquisition. And again, I'm very excited about all the ancillary technologies because they're hot, they're closer to my area of expertise and passion. So it's all gravy. Speaking of one of those fringe technologies, object, you got an object play and bycast and what's going on there? I mean, what do you see as the potential for object? So I chair, as some of you might know, the Cloud Storage Initiative, which is basically the marketing and education arm of the technical working group in charge of the first cloud standard, which SNEA happens to have released, that's the Cloud Data Manager interface. That is all about an object interface to storage that's completely cloud optimized, very cloud friendly. The benefits of object are back to that one characteristic that distinguishes this new generation of cloud scale. What, you know, it's heretical for a file systems company like NetApp to say this, but the truth is as you get to exascale of data, as you get to many billions and trillions of individual objects in organizations have to manage, you just, you're exceeding the design goals of POSIX compliance file systems designed 30 years ago. Having to structure things in folders, hierarchical folders, having to look up, you know, file names and open files and seek to a certain region of a file, that's just way too much overhead for where this world is already here and going. So an object-based interface lets you abstract, you know, the location of the file of the object from, you know, how you're accessing it and what the application of the user needs gives you that flexibility to have, you know, either a flat namespace of billions or trillions of objects without having to have folders or it just says, you know what, I care about the data much more than who last accessed it or what time or who owns it. I need to know how many people access it. I need to have legal polls associated with that file. I need to have other demographic information associated with that object. And I don't want to have that spread across three or four different databases that might become inconsistent over time. I want the metadata tagged to the object directly. That's the freedom of flexibility that object storage offers you and standards like CDMI are definitely catalysts for driving more rapid adoption. Yeah, okay, so what's the play there? Are you, does ONTAP speak object? Is that where this goes? So bycast, the bycast acquisition, which is now the storage grid product line, NetApp storage grid is all about gaining all the expertise in terms of the management layer and the interface layer for object storage. We obviously have a lot of core competency in just storing lots of bits of data very reliably, very economically, very rapidly. What we don't have a lot of experience in is object interfaces for doing that. So we know NFS cold, we know fiber channel cold, iSCSI, SIFS, you know, but emerging object interfaces which have been proprietary in the past because it's an API until the standard emerged. Certainly the most credible, viable standard is the CDMI standard. We basically, without having announced, I think a product yet, Mizeo Software did that on Monday, so I was very proud to see that. Certainly it's cherry in the CSI, but we have all sorts of intentions of making sure that this standard, which we're strategically investing in, is the object interface standard. We want to be a dominant player in object storage using this interface. So I have a question for you, just to kind of change gears a little bit. Obviously you're in Silicon Valley, NetApp's known to be a very big success story for many, many in Silicon Valley, including the investors. Ottawa. Like Sequoia. And we follow you on Twitter, so we see your patterns. You're scouring the web. You're actually doing research and going to events. What are the hot areas that you see within this economy that we're living in, in terms of from a startup standpoint and from emerging technology? I mean we haven't talked about automation. We hear a lot from developers that in the automation challenge with this new migration, this new transformation going on, is a hot area. What areas do you see? I mean you're out in the hinterlands of the tech land. What are you seeing that's hot these days? I love these areas. Obviously big data. I'll give you one that I'm looking at, in fact, actively researching very, very closely and it's a bit of an emerging one. You might not have heard of it yet or hopefully you have. And it's more of a trend and it's a category of technologies called DevOps. One word. Not sure if you guys have heard of it yet or if you hear a chatter about it. I'd love to know if you hear a chatter about it at this show because that'd be particularly interesting. Exactly. I mean I think the spring source, that open source crowd where you have that infrastructure automation, the right scale kind of. But I think the maturity basically of fusing two very separately minded organizations, people that are paid to resist change, operations, and people that thrive on change, application developers, having those two cultures merge, get together and apply agile techniques, not just the development which is the most progressive way to write software today but the infrastructure. Managing your infrastructure programmatically, version controlling your infrastructure, applying all sorts of dependency rules so that you can get as complex and fine grained as you want and still manage the firmware on a million disk drives on a thousand Hadoop nodes at scale. Being able to do that without humans involved but just programming that. That is a cool. We have some programmable infrastructure. Absolutely. We actually are following that and we don't talk about it because it's not a real market for that but what you're talking about is the pure play cloud guys, right? Guys who have no legacy. All the young entrepreneurs that we're seeing with DevOps is they got a clean sheet of paper. They have no legacy baggage. They go, hey, I'm a coder. What's this? Amazon. This is so easy. I can pay on a credit card but they get Amazon as a programming environment. Not as an operational IT. In my mind, Amazon is the leading example of DevOps and currently it's an organization. They've always programmed their infrastructure. They've always used it maybe initially for efficiency but they quickly tossed that like early last decade and focused on agility and ability to deliver innovation more quickly. My vision on this is that DevOps is going to emerge into, and I haven't written blog posts about it yet but since you brought it up, I'll get on the record and probably be correct in the future. Is that, you're talking about programming infrastructure. Yes, I am. I mean that's exactly what this is. This is not about IT and plumbing. It's about programming infrastructure. Well, it's going to have to be about IT, right? Because this is where IT inevitably is going. Particularly IT operations. And so if you don't recognize this trend early enough, if you don't start to reach out and understand your developers, if you don't have some kind of comfort in programming and infrastructure versus just operating it, you're going to be left behind sooner or later. And the clash of the Titans in this battle is going to be can the ops guys let go of their fiefdoms? Whether it's Fiber Channel, you would mention that debate was going on in your session. The programmers just want things coded and done, right? So you've got a developer mindset and the ops mindset. I mean they're completely different mindsets and don't you agree? Yeah, I agree with your points. The developers are driving this naturally but the developers aren't being naive about this. It's not about forcing IT operations to cater to their will. It's about developers realizing that operations is hard and actually writing better, more fault tolerant applications, more self aware applications, more instrumented applications. So it is a bit of a give and take. It's not one sided, which is what excites me. And it's a young movement. It's a very early movement. It's a global movement because it started in Europe and Brussels actually and it's spreading over now and becoming quite topical here. So I think- And there's some successes too. Look at Heruku, a great example, sold to Salesforce. Exactly. These kinds of environments are out there. I'm a big believer in this as the future of IT. I totally agree with you. Totally agree with you. How you scale applications in the cloud and operations of the scalable apps in the cloud. Yeah, I totally agree with you. I think you're 100% correct on that one. The challenge will be in IT is how fast does it infiltrate the enterprise? Cultural and organization will be the challenge. So this might be where the online service providers become the gateway infrastructure for those guys. Because if you're a programmer, you're going to go where it's frictionless, right? Either you do it yourself. There are sea changes every decade or so in this industry and I predict this is going to be a sea change for IT ops. Enterprise application developers actually went through a really rough phase 10 years ago when a lot of that work was outsourced overseas. And they've kind of come to an equilibrium now where they're business analysts and certainly innovators now in terms of big data, value to a company, data-driven value. That sea change now is being forced upon IT operations. I said in my breakout, to be melodramatic just so it's stuck, that ultimately cloud is nothing more than a massive assault on IT operations by developers. If you really think about it too, it's like an end game over time. And you have to sort of wake up to that reality and adapt. If they can do it at scale, I think this is where I think the lever will ultimately push that and just kind of pregnant skating is that if the scale, you brought up scale, because I think that's the pivot point. If scale change can happen at scale, then you've got success. That's where the winner will come from. The economics don't work if it doesn't work at scale and the economics ultimately trump up. So it's cloud the final resting place for IT operations? Is that? As far as we can see, and of course, things will change, come 2020. In fact, I recommend you spend some time looking at the published works recently of Bill Coleman, founder of BA, father of Solaris and so forth. He's at a VC now. And I had the pleasure of looking or participating in one of his presentations where he spoke about the evolution of cloud and pretty much decade boundaries. Cloud one, as we saw ending in 2010, more or less the peak of innovation. Cloud two really in this decade. And he already had cloud three laid out quite eloquently starting in 2020, 2030. And even some of the implications on humanity beyond that, when you have all sorts of computing available as a complete utility. So I think it's fascinating where this is going and there will be something beyond the end of IT. Well, let's not stop here. We're in Palo Alto. Want to get you more involved in some of these conversations. One DevOps one, I think is that's a franchise topic right there. We could love to pound that home. I'm interested in that. Obviously Hadoop, the big data stuff. And also just the change of the infrastructure a little because this is going to be a massive fun run for IT and tech guys. It is. I mean, it's already very frothy investment-wise. Change brings opportunity. People are going to make a lot of money and good stuff's going to happen, right? They'll be some dead bodies, but it's going to be fun. There will be an open source angle here that will loom larger than ever before in commercial solutions, money-making ventures. So that angle too will be fascinating to watch. Yeah, I mean people, I constantly ask the question, how do you make money at open source? Well, we're seeing it all around. Of course you are. And it's interesting. It's refreshing to hear an executive with a storage company talking about some of these trends. You typically don't hear that from the storage world. You don't hear a lot of it at the show. That's why I was really, really happy, Val, that you came on the show. And we've got some startups coming on too. We'll hear from them. That's where we get a lot of good content from the startups, because they're on the total front lines. My biggest value of all these shows literally is to network and meet these interesting emerging technologies and startups. That's, you know, I know our sales and marketing department love the sales leads. I selfishly go, you know, I go for the networking. Yeah, yeah, yeah. Get the content. They're coming in the cube. We really appreciate it. We've had you on our list for a long time. Target list. Val Bercovici, great prognosticator, trend spotter. Twitter aficionado, you're very candid. Really appreciate it. Thank you so much. Appreciate it. Really appreciate it being on here. Cheers. All right, Dave.