 At Big Data SV 2014 is brought to you by headline sponsors WAN Disco. We make Hadoop Invincible and Actian, accelerating Big Data 2.0. Okay, we're back here live in Silicon Valley, this is Big Data Silicon Valley. Big Data SV event, SiliconANGLE and Wikibon's the Cube, our flagship program. We go out to the events, extract the signal from the noise. I'm John Furrier, the founder of SiliconANGLE, Joe Mykoz, Dave Vellante. Co-founder of Wikibon.org, our next guest is Jack Norr, CMO of MapR. It's a lot of news here with MapR, Cube alumni been on many times, big supporter of the Cube. Great to see you again. Thanks. Now we're interviewing you, we'll go in that story later. Yeah, why don't you guys start asking us questions. Yesterday, you guys had a great party we swung by and you actually interviewed the Cube guys. That was a thrill of a lifetime to be on the other side of the camera. I gotta say, I was really nervous. Dave, how are you? I was feeling pretty good actually. It was fun. Well, welcome back to the Cube. It was all fun, we can always joke around that. We have great conversations, we always do. But you guys, again, the original three, we call it the big three of Big Data, Cloud Air Hortonworks MapR, not necessarily in that order. You guys have different approaches all growing, doing well. Big news for you guys, one, rebranding and two, a lot of news around traction. But also you did a partnership with HP. Yes. So expand on that. What does that all mean? Why the rebranding? Why the partnership with HP? What's the update? And we also announced a sandbox available. So it's a fast on ramp to Hadoop. And we also announced our yarn support, which I can go into detail. And I'm going to save the details on the HP Vertica integration with MapR. You've got a couple of important guests coming up soon on the Cube. So I'll let them break all the details. Okay, call Mahoney, he'll be coming on shortly, we hope. I think he's landing. He should be on his way from the airport, but we'll hold that for him. Yep, and our CEO, John Schroeder. It's pretty big news. It's actually, the response there has been dramatic. And it's another proof point about the power of the MapR platform. So we've combined the power of Hadoop with this very flexible data platform. And that flexible data platform that provides full random read-write and provides an interface of NFS has a lot to do with the advantages with yarn. I was talking to John last night, and he had a spring in his step. And we were talking about all the companies that we didn't recognize on the show floor, and we were kind of like rolling our eyes. Wow, a lot of new companies hitting the floor. Some might not make it. Some might be new entrance of the market, which is pretty getting pretty crowded. But he had a comment when I asked him about how things are, and he made a comment. I thought it was interesting, I'll bring it up here and have you expand on it. And he says, people in business like a business model. And you guys have a good business model. You have licensed revenue. It's known, tried and true, successful business model. Different than the others. That's a good or bad, just different approach. And he's got his head a little smile on his face. So things seem to be going well. So could you expand on what he means by that? And what does that mean for the marketplace and your efforts? Yeah, it's a proven business model of really focusing on a product, selling a product, making a product enterprise grade, combining the innovations of the community. And we're very active in the community. It's a robust community. But providing some innovations and advantages so customers can be even more successful. There was a panel this afternoon, actually earlier this morning, where we had Cisco and Solutionary and the Rubicon Project. And the Climate Corporation share their production stories of Hadoop. And the Cisco representative talked about, within Cisco, they focus on how can they best leverage the data. They don't focus on what are the issues around Hadoop. They've got a platform that works. They focused on the architecture at the beginning. And now they're really dramatically expanding their use cases and how to drive value in the organization. And I think that's what a business model does. It helps an organization focus on their business and how they can drive value out of their data. Well, we've been hearing all week that a lot of suits at this strata, more so than the classical hoodie crowd. So with suits come wallets and with wallets, it's all about the ROI. So let's have that ROI discussion. We're sort of entering the face. We've been commenting. A lot of the discussion here is about integration, partnerships. Really maturing the ecosystem. And to me, that's a signal that now we're starting to make some real money here. We've always put forward the premise that practitioners are going to create more value than the supply side. And that's your first Hadoop application might not return as much as your 10th, but it's probably a really steep curve. So what are you seeing in the ROI? What's the discussion like with the CIO community? Well, first off, the big rocks, the big picture are Hadoop can really drive the top line. So we're seeing applications that drive upsell and cross-sell and new revenue opportunities. And there's a whole class there. Sometimes it's leveraging new data sources. Sometimes it's just expanding the data sources that they have available. Then there's a class of applications that are around risk mitigation. How do I detect and prevent fraud? How do I really understand my risk exposure, whether it's portfolio? In the case of Climate Corp. They were talking about how do they develop new insurance products to address the major issue with agriculture and 85% of the risk there is weather-related. So that's a whole bucket. And then there's a bucket on the operation side. And a lot of that is how do you drive efficiency, whether it's supply chain or whether it's detecting failure and preventing downtime, or just straight cost? And in terms of I'm offloading data warehouse, I'm looking at really a disruptive force of Hadoop, which is delivering the ability to store, manage, and process data at a dramatically lower cost. We're talking 150th the cost, 120th the cost, depending on the platforms you're using. So I want to ask you specifically a question about fraud and even credit card theft and bring it back to the consumer, the everyday person that might be watching. So last, I want to say two out of the last maybe four Saturdays, I woke up, I got a call from American Express saying, did you make a charge in Indiana for some electronics store? No, for $2,000, no. And so you're seeing real time. I mean, that morning they made the charge within minutes they were calling me. Is that Hadoop behind that technology? It's got to be big data. Are you seeing those types of applications? That's an example of how we, that used to take six months. I don't want to get into company specifics, but in general, credit card, financial transaction, absolutely Hadoop is being used for that. And it's being used so that you receive the call. It's also being used so you don't receive the call when it's a transaction that you make, maybe not very frequently, maybe it's every six month pattern, but they can determine that. So that's Dave that did purchase that. Right, right. And then you're talking about risk as well. So we're talking here about risk profiles. Do I loan somebody the money? Do I underwrite that policy? Do I actually maybe cut the price and be more competitive? You're seeing those types of real time decisions as well. Yeah, it's a very horizontal uses that we're seeing. And what's surprising to me, it's not just a cross industry we're seeing within a single company. You're seeing applications across that company, 10, 20, 30 applications running directly on the platform. And there, it's not just about is the platform enterprise great and can you provide the SLAs and data protection, but it's how do you integrate that with the existing environment? How do you provide multi-tenancy so this workload doesn't interrupt this workload and this group can only see their data? A lot of security around it, a lot of operations, a lot of it just making sure this thing runs because as you add more and more applications on top, by the very overlap it becomes mission critical. And that's kind of what I was alluding to before and I like the way you describe it. But John and I always talk about the sort of, the return curve is really steep and exponential almost. Like I was saying, the first app maybe spend a dollar, you might get 1.2. And now you're seeing as you add more and more apps, you call it upsell and cross-sell, in particular that cross-sell piece, you must start to get much more operating leverage out of your Hadoop infrastructure and as you permeate other parts of the organization. So what are people seeing in terms of that return curve? Really steep like that? Well, I don't know if I would characterize it as flat at the beginning, just because that happens to be the first application. I think it is true that it is exponential on some of those use cases that might involve advanced algorithms where you're really moving the needle in certain applications. But it's pretty easy to identify low hanging fruit and say, look our first application, let's not do something with exotic machine learning, let's deploy Hadoop, it's going to be this enterprise data hub in the organization and let's offload some of the data, let's offload some of the processing from very expensive data warehouses, very expensive enterprise storage, very expensive other applications. I'm going to house it there, I can process and I can selectively upload the data. And the beauty about that is as I add more and more data, I don't have to pull large chunks of that data and process elsewhere, I can do analytics and sift through the data and find exactly what I need. So it's actually even more usable by storing it in Hadoop. So what's your take on what's happening to the traditional enterprise data warehouse? How is that being complimented, supplanted, replaced, enhanced by what's happening with Hadoop? I don't think organizations are saying now that I have a technology that can potentially replace a data warehouse and I'm going to rush to replace a data warehouse. I mean, if it's working and you're getting value out of it and you're getting applications that are downstream, keep using it. There's so many other places to deploy Hadoop. But what we are saying is that just because I've got a data warehouse here and my data is growing by 40% a year, does that mean I need to invest 40% more in the same infrastructure? There's no alignment that if I do that. If I do that, there are some efficiency where there's huge stair steps and it doesn't necessarily translate into huge stair steps in terms of the business value. I think that's where Hadoop fits in. I can deploy Hadoop. I can offload some of that data growth so I can take my existing infrastructure and I've got a cost lever now to say, well, hey, data doubled, but it doesn't necessarily mean that my data warehouse has to double. And I can do the processing on the platform. We have a customer that was a large telecom provider. They were doing a lot of complex transformations in the data warehouse because it couldn't be done outside of it. Well, now they can do that on Hadoop and just upload the selective content. They also had a lot of cold data in the data warehouse. Well, they could pull that data off and load it back if it's necessary or do processing on the Hadoop platform. And I think that makes a lot more sense. It's not necessarily a rip and replace, but you can get enormous value just by augmenting and offloading. Yeah, it's not a rip and replace. It's a, John, we joke sometimes, it's a huge sucking sound. Now, you were talking about yarn before. Talk about what yarn has meant to your business. How are you exploiting yarn and what's going on? Yarn's very exciting. Community-driven effort to basically expand what's possible on Hadoop and not use just a map-reduced framework. Why we're particularly excited about yarn on MapR is we're combining this general-purpose compute with a general-purpose storage. And what I mean by general-purpose storage, it's random read-write, it works like enterprise storage. So if you've got an application that was written for, let's say it's an HPC environment and it's using a standard file interface, well, you can take that and run it on the MapR platform and it works without changes. And that's very different than, I'm gonna run it against another distribution and then have to understand the limitations of a write-one storage layer that's underpinning Hadoop and introduce a file close, et cetera. Let's talk about SQL on Hadoop. So we just, Rishi from InfoObjects was just on. He said that, I'll put it in words, we've seen consolidation in the no-SQL market spaces. We're down to 200 across the street. But it seems like, and I was tweeting with Ray Wang the other day, how we see SQL growing as well as no SQL. So what are you seeing in terms of SQL on Hadoop? Basically, there's a lot of activity, there's a lot of interest. And it's basically, I've got this data that's available on Hadoop, doing some important processing, but I wanna share that across the organization. I wanna share that with analysts that maybe aren't Java-level programmers, but are SQL jockeys and understand how to use that, wanna use that against the Hadoop data. And that's why we're excited about relationships like Vertica that's 100% compliant and works like you would expect it to work. What are you seeing for, what are you thinking about some of the big questions that you have going forward in this marketplace? I mean, you guys have been laser focused. We talked about last night on the video we made with you, that you didn't really worry about all the noise. You just said, all right, we're gonna go solve a problem. And that's what we've always liked about MapR. But when you think about how this market evolves, you've seen a couple of pretty strong IPOs and guys like, well, Splunk for example, and Tableau sort of, I mean, more visualization, not that sort of core Hadoop space. You're starting to see guys like yourselves and some of your competitors start to get to critical mass. A lot of people talking about IPOs, we always pontificate about acquisitions. You're seeing the big whales participate very actively in this market, certainly from a marketing standpoint. And even if the marketing leads the product, they'll eventually get there. How do you see this market developing and shaking out? Do you expect that you're gonna see several large pure play Hadoop players emerge? Do you think they'll get subsumed by the big guys? Do you even think about that? So I think the starting point is this is probably one of the biggest changes to enterprise software that we've seen in quite a while, certainly my career. And that's driving a lot of this opportunity. And our focus is not just an incremental improvement in data warehouse environment. You've got enterprise storage that's separated by compute. And really this is about a distributed framework that has those together and takes advantage of that processing across a massive cluster. You can do so much more in a much faster period of time. We also have a separated data warehouse environment from production. And really that was not an original design. It was basically production, couldn't handle the analytic workload and bringing that together and doing more operation analytics where you can take high arrival rate data and do analytics and adjust to business case, whether it's detecting fraud fast or doing recommendation engines or detecting sensor data and improving operations for the business. All those have a huge impact. And that's where we see a lot of the move the needle type of applications. And that's our focus that the whole theme of RM7 was how do you get an enterprise grade database functionality into the Hadoop platform so that you can do things side by side and do things in a real time environment. So John's, I'm dominating the conversation here. I'll give up the mic in a second. But obviously large companies can innovate, Apple innovates, you certainly Google innovates, but large enterprise companies, the market you sell into, there's not a lot of innovation going on there. There's a lot of business that's done, a lot of acquisitions that go on. So I ask you, can large oligopoly, can those companies innovate or do they need to partner with, acquire startups, companies like yours and others? I think this is, I mean, this is an architectural change. This isn't, let's take our existing product line and do a few tweaks and now we have a big data platform. This is an architectural innovation and I think that's why we're seeing such a wide open, fast acceleration of companies like MapR. So, welcome, I'm glad I'm back, I guess I've worked on. Could you introduce yourself, it's been so long. Jack, so I gotta ask you, obviously we had a great time last night, you guys had a lot of customers here, met a lot of folks in the industry too, great event you had here at the Hilton. Obviously the social proof is there, MapR is on the map and on the map, has been on the map. So I gotta ask you a question, going back to the beginning of time of MapR's founding and when John and team put it together, a lot's happened. The early critiques were out there and we were on the record saying, you know, I was telling your board member, Pete's on CNA over email, we were there supporting you guys at the beginning saying you watch the enterprise and if you long they stay on the enterprise track it's not gonna hurt anyone. We always said that was a good thing. So I want you to go back to the beginning of MapR and look where you guys are now and the question is what has surprised you that has happened since and what hasn't surprised you? In other words, what happened that you thought would happen and what didn't happen that you thought would happen? So what was the surprises and what didn't surprise you going from the beginning as this evolution evolved on the big data market? Well I have to give credit to John Schroeder and MC Srevis, you know they, when they started the company it wasn't clear that Hadoop was gonna be the winner. There was a lot of big data technologies out there, they identified Hadoop as the platform and they identified this need for kind of operational analytics. If you go back to the original business plan and presentation it's been a pretty laser-like execution. So that hasn't been surprising. Maybe the surprise has been the speed and development of this ecosystem. You know the number of companies that you don't really recognize that are on the floor that have emerged and the focus that this is the topic of Wall Street Journal articles and articles across publications on the importance of big data and having the importance of a big data strategy. And we thought that would happen, I'm not sure if it was beginning of 2014 that we'd quite see that activity in momentum. Well what surprised you in terms of the lead is remember the early days, oh this distribution for that distribution and MAPI was going what are they doing? So like it was a big argument that never really kind of panned out to be relevant. I know you guys were saying that as messaging but a lot of people were talking about that. Did that surprise you, that that happened quickly sooner than later or did it drag on longer? What other things can you share? So can you state that again? Well remember in the early days it's like oh, Cloudera was there and then Hortonworks came out. Then these sort of articles like on Gigong, oh the battle of the distros, who contributing more to open source. And then this MAPI are doing their thing over there. Then it'll be okay, there's another approach. So kind of the war of the distros was a big discussion. Turned out to be kind of irrelevant when you look at it now. Did that surprise you, did you guys see that coming and did it go on longer than you thought or did it kind of go out faster? No I think when you say open source people automatically think well there's one business model. Most of the open source examples out there are the commoditization stage of a life cycle. And here we're at the beginning with Hadoop. So I think initially there was this kind of confusion and I think the hybrid model that we have where you take an active part of the community and you have open source that's part of your distribution and you're combining that with innovations so that it's easier for enterprises to get to value much faster and to support a broader set of applications. There's been a pretty fast and broad acceptance to that with the enterprise customers. So what's next for you guys? Let's kind of look forward, okay. Post strata, post a big data SV event. What's on the plan for next year? Next six months, we'll say six months. What's the tactical plan? What are you guys looking to knock down? We are consumed with helping customers move into production. I think we're, you know, we've seen a lot of tests and development. We've shared with you some of the examples of production use case and when you got customers who are doing 90 billion events a day on the platform, we've proven out our production use and we're just focused on helping more and more customers move into production, do that seamlessly and expand their footprint use of Hadoop. You feel good about the party last night? Good time? It was good and I must say the highlight for those of you kind of, when we had our video crew and I was able to ask you a bunch of questions that was, you guys were good sports. So let's talk about this. So what happened last night? We went to the MAPR party, which is next door to our big data SV event here. We have our party tonight from six to 10. So if you're watching here at the strata conference or in Silicon Valley, come down and visit us and come visit six to nine. You had a little film room. So you asked us to interview, you interviewed myself, Dave and Jeff Kelly. We're standing there. You were the, asking the questions. You were, you had your own little cube operation there, which was great. And like the cube, we don't tell people what we're going to talk about, you know, really. We didn't know what you were going to ask. So, so John said, you said you were a little nervous. Of course, very nervous. I'm nervous right now. Nervous every time I lay down. No, no, it's going to, what I'm going to say, you could kill me. We'd love the cube. Thanks for your support, Jack. You guys are great. It's fun working with you guys. You know, and I got to say watching the industry grow up and seeing the early players there from the beginning and start to, you know, reap some of the fruit off the tree, especially with all the new entrants coming in, it's good to see it's mature. As Dave always, Dave and I always say, you know, we're a president of creation at the Hadoop trend of big data and it's fun to watch it grow up and be part of that ride. So we love it and we love building the relationships. That's what the cube's all about. Really appreciate it. And again, we'd love to continue to do it. We found out yesterday that we interviewed over 3,000 people, 3,204 people before we started this event. So, you know, it's been great. We do a lot of talking, John. Thanks for your support. We'll be right back after this short break. And we're going to have the CEO of MapR coming on later today with HP, that big news. Stay tuned. We'll be right back.