 Okay, we're back live at Strata. This is siliconangle.com, siliconangle.tv's the cube. I'm John Furrier, I'm with Dave Vellante from Wikibon and we're breaking down the analysis and it's a ton of action going on. O'Reilly has live feeds on their site from keynote sessions. They have obviously great content, we love O'Reilly and they have also tons of live streaming. They have their little round table, kind of talking about the books, talking about how-tos, some great live streaming here. It's all about the analysis and I'm pleased to welcome back on theCUBE Mike Olson, the CEO of Cloudera. Cloudera is close to our heart because they've been great partners for us at Siliconangle and theCUBE. We were in their space for a year and a half, got to know the team over there. They're doing extremely well and we love those guys. So they're friends of us and ours and they collaborate really well and if their collaboration with us has been a big part of our success, I want to thank Cloudera for all your support. Mike, we can be thankful all day long but you're thanking the company right now, your team, because you guys are kicking some serious butt right now. Great product, great team, market's hot. I mean, literally here just like a Duke World strata is super busy. Tell us how you feel and what's happening. Well, a few things. First, absolutely mutual. We've enjoyed all the time that we spent working with you, both alongside and watching this market unfold. But you're absolutely right. The market is exploding right now. I think things are going extraordinarily well at Cloudera. I think in part that's because we're executing pretty well. But in addition, the pace of growth of the market and the intense interest that we're seeing in big data solutions is just overwhelming. If you walk the show floor here at Strata, you see large established companies. You see smaller businesses you've probably never heard of before that are building some really compelling and interesting applications. It's an exciting time in the market. I think this year, next year are going to be really transformative for the way that data is managed by business. I know Dave Vellante wants to talk about the market sizing that Jeff Kelly and his team put together, the first groundbreaking market sizing market data report out there from wikibon.org. We'll get that in a minute. But I wanted to talk to you about kind of the evolution of how this marketplace is exploding. I wrote a couple blog posts this week, essentially comparing big data and all the things behind it, all the disruptors like flash and virtualization, things that you know a lot about, as being a new industry like the PC did to the mainframes, data's unearthing a massive amount of jewels and opportunity for entrepreneurs in business. Dave and I were talking yesterday that 2010 was the what is Hadoop year, and obviously you guys were well in front of that. Big data as Al Marawadala said, you know, you saw the future years ago. And 2011 was, hey, big data's a real viable market opportunity. 2012, the platform is now matured, Hadoop's an industry standard. And as you mentioned at your last keynote at Hadoop World, applications are starting to hit the scene. We're speculating that 2013 will be the year of payoff, beginning to see the reigning of value in dollars and benefits. So do you agree with that one? And could you add to it and maybe add some color to that vision and share with us your thoughts on where we are and kind of what's going to happen? Yeah, you bet. I absolutely agree. If you look backward at what's happened over the past several years, a few really fundamental trends have driven, I think, the market forward certainly created an enormous opportunity for Cladera. Number one, the widespread availability of data that just wasn't there before. Machine-generated data from devices talking to devices, much more consumer and individual-generated data off of social feeds and others, it is now possible to collect information in a volume and in a variety that wasn't available before. Given that that is all new, given that there is so much of it, my contention, and we see this in our install base, even relatively dumb ideas, even relatively simple analytics pay enormous value. None of the fruit on the low branches of the tree has yet been picked. We simply haven't been able to get there before. So data availability is driving business opportunity in a fundamental way. You know, you said something interesting though early and that is that flash and hardware and other trends are driving the market forward. Let's be honest, enterprise software and enterprise infrastructure have been boring for about 15 years with the emergence of widespread inexpensive available compute with a dramatic reduction in the cost of huge amounts of storage. But then, in addition, driven by some of the advances that you're seeing in flash memory and in high performance wide area networking and so on, I think we're watching a new platform, a new infrastructure evolve that's going to pay dividends in ways that will surprise us five years from now. Yeah, let's drill down on that. That's a great point. And we've had some other guests on earlier yesterday from some big companies. You guys have been complimented kind of off the record from some of these bigger companies. I won't say their names. They have three letters in the acronym. So these are a few you can pick from there. Complimenting Cloudera's consulting efforts and how effective you've been in the field. So not only have you been a really successful startup with commodity, using commodity hardware, but you've actually been very successful talking to customers, right? Second thing is we just talked with Scott and we were talking yesterday about this new systems architecture that flash at the server has been causing and the data centers are still in the early days of flash. So let's kind of look at that market that you're talking about in the data center and in the, let's take out the big consumer comes like the Facebooks and whatnot. And they're kicking ass and they have their own little world that they operate in. But the normal big enterprises, right? Enterprise software. They have virtualization. They've had for many years VMware has been dominant. Flash is hitting the scene hard and still in the early days. Now the big data message is so disruptive. They're all kind of kicking the tires. They need help. I posted this morning on Forbes that the side effect of the big data movement is a services boom in the consulting area as well as the product side, whether it's SaaS and or actual product. But the real elephant in the room, no pun intended, is the fact that the data warehousing in the business intelligence market is old legacy software that customers have paid millions for, right? So they need to do something. They're not going to just throw it away. Yet Hadoop is a green field opportunity, hence just a free clean sheet paper to work from. But you also have to compete in these environments where there's existing stuff. How is it exciting and what, share with the audience kind of what's the state of that market with the BI, the data warehousing? I'm an old school relational database guy. I worked at Illustra and at Informix. I built a database company of my own. I sold it to Oracle and I spent some years there. I believe deeply in the power and in the value of those legacy solutions. Look, relational databases evolved over 30 years to solve very important business problems and they are fantastically good at attacking those problems. If you need to do online transaction processing, online analytic processing, I'd advise you, go talk to IBM, go talk to Oracle, go talk to Microsoft. They're the market leaders. They have wonderful product. Hadoop is complimentary. It offers new capabilities that were simply not available prior. It allows people to answer questions that were previously unaskable. But it is only useful if you can stand it up in your enterprise data center, if you can integrate it with the rest of the infrastructure you have. Now, you're right. We've got a thriving services business, both helping customers build applications and also thinking about how to deploy this new platform. But fundamentally, we're a product company and a big part of the value we deliver is manageability, monitorability, administration of Hadoop and production and integration with all of that legacy infrastructure. How do you talk to the BI guys that have those investments? Because they all want to have a Hadoop component of it. Is it, you have to integrate? Is it, do you put a wall around it? Do you wrap around it? Do you integrate with it? What are the fundamental challenges that people are, what kind of questions are they asking themselves that you can shed some light into? So many of these folks, Tableau and Taland and Pentaho and others have built integration already with Hadoop. We've worked very hard to deliver an ODBC driver that makes that easier to do. So you can take your existing user interfaces, your existing analytical tools and you can point them at a Hadoop cluster. That's great, that's a fantastic first step. But really I think the opportunity is to build out new visualization tools and new analytical tools that know that there's complex data and massively parallel processing infrastructure under the covers. And I think we'll see new but also established BI vendors take that opportunity on full board. So I buy Mike that it's, that Hadoop is complimentary, no doubt. But there's something else happening here which is that Hadoop infrastructure, which maybe today is whatever. There's some small percentage of the total is getting more and more and more important. And do you see that sort of dynamics flipping where no longer will the tail be wagging the dog where Hadoop becomes the dog? I think in terms of raw bytes stored indeed, the majority of the world's data over the next 10 years is going to move into Hadoop. Most data will land there and will live there. It'll be even more important to integrate. That is to be sure that you can move the data from whatever platform it's in. Hadoop, if that is the repository record into the specialist systems for that kind of analysis. But yeah, absolutely. The big untapped opportunity here is for companies to stop throwing away detailed data that's valuable, but that's too expensive to store right now. We've driven the platform price way, way down. We've made it much more possible to analyze that stuff and we'll see new infrastructures emerge. So it's a proxy for value, the raw bytes stored. Is that fair? Yeah, and in fact, when we talk to customers, one of the terms we use is return on byte. The amount of value you extract from a byte of data stored has to be greater than the cost to store it. The fundamental innovation of Hadoop is it's driven the cost so low that it's feasible, it's affordable to store stuff you used to have to throw away for cost reasons. What about, can you share with us, because you're just talking, you love the product side. You've been an inventor, entrepreneur, CEO, and you know the database world. Talk about the impact and the massive surge in HBase. Why is HBase, I mean, we know why. We're doing an error on HBase project we haven't talked about publicly, but you know. But in general, you've got the HBase conference that you guys are funding and promoting and you're going to have in a forum for developers. Why is HBase just taking over as a phenom right now? What is the, what's the angle there? Well, a couple things are important. One is that because it is built on because it's fully integrated with Hadoop, you get the full faith and credit of MapReduce. You get all that analytic power over your HBase data. It's also a very nice programmatic interface that is developers who want to build applications to put and get data out of the repository, have a very easy time doing that with HBase, and it runs at interactive, say, web speeds. So I can get past the limitation of Hadoop being batch mode only, HDFS, high latency. I can get interactive speed, data storage, and retrieval, and I don't have to sacrifice the power of the analytic platform underneath it. So, Mike, you were talking about, John was actually mentioning Hadoop, the Hadoop 101 conversation that we had in 2010. You were having it well before that, and now we're really talking about real business applications. Talk a little bit about what customers are doing, what they're telling you, and what are some of the barriers that you're seeing? Well, I've said this a bunch and it's important to highlight. I love the progress that we as a business have made, that the platform that big data have made in the market, but let's be honest, the gate to adoption right now is applications. A financial analyst, an insurance adjuster, a researcher at a hospital needs applications that understand his or her data that can ingest that stuff and provide meaningful analysis of it. Those apps are beginning to appear, but it's early days yet. It's sort of 1985 in the relational database market. We'll see lots more come into existence, but it's absolutely critical to us that people build on top of this platform in order to drive that value. Now, as to the business problems that people are tacking with the platform, there's a wide variety of analyses that our customers are doing. It depends very much on the vertical market in which they are. If I look broadly across the markets, lots of customer behavioral analytics, why do people do what they do? How can we predict what they're going to want from us in terms of content, in terms of advertising, in terms of new product and so on? A much better and deeper understanding of holdings. For example, the value of inequities or bond portfolio, a better assessment of the risk of my holdings is a common theme across the markets we look at. And then also really turbocharged, high-powered security infrastructure. I don't want to just monitor for penetrations and attacks. I want to keep all of the activity that happens on my network and I want to forensically analyze it after the fact. Looking for evidence of attacks I didn't recognize at the time so that I can strengthen my infrastructure and that is across industries as well. Is that security problem yours or your ecosystems, both? The platform that Hadoop is and CDH and Cladera Manager make that platform manageable, but that is the data storage, data processing infrastructure. The actual analytics, the actual threat assessment that runs on top of that, that's an app that we and our customers have built jointly to roll out for them. And again, I think it's an opportunity for people to build new value, a new shrink-wrapped application that did that would be fantastic. You guys have had a mission from day one about the vision of Cladera and it's funny in talking to anyone who works at Cladera, it's such a great culture. But as the market's maturing and you guys are growing super fast, has any of that mission changed? I'll see, you're not going to re-engineer the trajectory of the business. You can make some tweaks. So talk about the mission of Cladera. Has it staying on the same course, commodity hardware, open source, and then the messaging? We've seen some things like, what does it do? Now we're getting to applications. You do have to tweak a little bit about the company's messaging. You're growing up. What are some of the changes? Talk about the mission, what has or hasn't changed and tweaks along the way you are right now. From the very beginning and still today, our mission is to help organizations profit from all their data. And I can go through that. Each of the words there was carefully chosen. Help, look, we don't deploy the systems for you. We want to teach you how to use this platform well. Profit from the data. It's not just about lower cost. It's about a bigger top line. That is why people deploy and use this infrastructure and all their data. Not just the relational subset. Not just the document subset. But anything you can get your hands on, you ought to be able to combine it in an analysis. That's always been what we believed was important for us to do. Obviously, as the market has matured, as we've gotten more experience working with customers, we've refined our tactics a whole bunch. And you know, I would say a couple years ago, sure, we were a startup. Now, look, we're a small company. Still a small company, but no longer a real startup. You're not 34 people. You're like in the hundreds. Are you across 200 yet? Indeed, indeed. And that gives us the capacity to engage with and work with customers and with partners in a way we never could before. I'll say, Mike, so thanks for coming on theCUBE. We know you got to run. Dave, any final questions for Mike at this point? Well, I could really get into it. I know you said you got to run noon, but I got a zillion questions. But hey, it's always a pleasure. I'll look forward to coming back and I just want to say, yeah, I'm the guy in the T-shirt. All right, thank you very much. Mike Olson.