 Live from Midtown Manhattan. The Cube's live coverage of Big Data NYC. A silicon angled Wikibon production. Made possible by Hortonworks. We do Hadoop. And when does this go? Hadoop made invincible. And now your co-hosts, John Furrier and Dave Vellante. Okay, we're back here live in New York City for Big Data NYC. This is where all the action's happening. This is the Cube, our flagship program. We go out to the events, like the signal from the noise. I'm John Furrier, join Dave Vellante. We're covering Big Data NYC, which is Hadoop World Stratoconference, with all the action, all the news. And I want to thank Hortonworks and Wendisco for supporting us. Thanks so much. And our guest is the president of Hortonworks, Herb Kinnins. Thank you for coming on The Cube. Appreciate it. I had Sean earlier on from Hortonworks. I want to say thank you guys for supporting The Cube, the community. Certainly very happy. We had some tweets saying, you know, for the record, The Cube has been amazing. But thank you for so much for that. Absolutely happy to support. We think you guys do phenomenal work for the community. So we think it's fantastic support and we're happy to go do that. So appreciate the opportunity to be here as well. So Dave and I were talking with Sean and we love talking business models and tech and go anywhere. But you're running the business over there. You have the clear view of the landscape and you see all the piece on the chess board. Obviously Sean has no more strategies. Got the moves down. But you got to grow the business. You got to fundamentally build the business. In the classic cases, what's the business model? I love that question. What's your business model? I want to ask. So let's just get on the record. Dave and I want to have a quick discussion about the Hortonworks business model. And what is your business model? Obviously it hasn't changed, but let's explain to folks. And this is where, as much as right strategy is important, how you build the business, what you do at the end of the day, it's proven in top line in terms of how you're growing the business and how you can go scale it. So from a business model perspective, our model hasn't changed one bit from day one of the company, which is there's really three core things we're going to focus on, which is we believe that Hadoop is helping to replatform the modern date architecture. And if we do this right, we believe you can go build a business 100% in open source and go drive that business because an open community should be able to out innovate any individual company at any point in time, if you can gather the resources of the community. So we based it on open community, open source, open source development and business model, based it on working very closely with the partners and building an ecosystem. So you walk into any Fortune 1000 company and they have legacy assets and they want to know, can I fit this new architecture and slot it very easily with all my legacy assets and you want to make sure that works and we haven't saw the partnerships. And third, make sure Hadoop's viable and consumable by the enterprise, has all the enterprise class features. So if you look at that from a business model, say if we're going to approach it that way, the right model is that the end of the day, what we provide is we provide all our software for free and open source, we put it all out there and we provide a support business. We support our customers as they go through the adoption cycle and what they do. And that support business is very similar to you would think of any other software company that has a maintenance stream and how they support their business and what they do. So the marginal economics of that business is identical to any software business, is that right? Absolutely. No, of course the training piece is not, training scales with people. Training and consulting. There's diseconomies of scale. So in your model mix, Sean was saying your ideal mix is 80, 20%, subscription 20%. Ideally 80, 20 between support services to training and consulting is where we'd want to be. Let's say today we're probably closer to the 70, 30 range, but trending right to where we want to be. And the reason for that is for companies to get successful adoption of a dupe, they need some help. There's a shortage of skills out there and they want some help. So we help them through that curve and make sure that they can get successful. Okay, so the profitability model, as we just determined, like any software business with a 70, 30 mix of software to services. But the argument could be that, okay, what you're giving up the upfront portion, the upfront license portion, chopping that off so you're shrinking your market size. How do you respond to that? So a couple of ways. So if you think about it, how do you want to grow the business? If you want to make a market function, what you need to do is make a dupe, a prevalent platform that can be used across all companies and across all partners and standardize on a standard distribution. You could argue markets in the past even in the open source world, they can either fracture upfront, which means different companies try to take on components of that, fracture the market and keep their market share. Or if you want the market to function and really all companies to be able to profit from this and be able to get advantage from this, you want her dupe to be an enterprise class platform that can scale horizontally, which means if you get everyone to start using this as the standard and you can go scale the business and you allow all of your partners to make money on top and to grow their business on top on it, there's incredible volume of how a dupe can become that standard platform. So you actually can make it up in volume, I was just joking, hey, we'll make it up in volume, we'll cut the price, we'll make it up in volume, you can't do that in the hardware business, you can actually do it in your business model. Absolutely, so think about successful companies who've done that on the software side and open source like Red Hat, right? You absolutely can build that type of model and make it successful and profitable. So what's your, you guys aren't afraid to talk about lock in because we're all open source, I'm trying to have a discussion about lock in, there's nobody less locked in than guys like you in Red Hat, but to make money you've got to have some kind of lock in. Even Google search, the lock in is that they've got great search. A VMware has a much stronger lock in or an Oracle of course has the mother of all lock in. So what's your lock in? So I wouldn't really say there's lock in because the way we look at it and say we do a one year subscription for support services that we provide to a customer and at the end of the year, they're free to say I would no longer like to do that. So your lock in is just doing a good job. They're in the right to get a renewal. That's your lock in. You are exactly right. So it's similar to Google's lock in. And what does that do? It actually balances the power between customer and supplier to say you have to continue to deliver, to continue to earn the right to do business. These are interesting discussions to me because you've got a monetization model, Google's got a monetization model, Java didn't really have a monetization model, right? And again, Google's is great search and they monetize it with advertising here. Android is Android is another interesting one. So open source to us is just so disruptive and unpredictable, but one thing that you know is that it drives innovation. It drives innovation. The other thing I would say is the market conditions are different now than they were even five years ago to allow a pure open source company to prosper. I mean, if you think about it, think about the maturity of open source in the market and how it's used today. It's companies are comfortable using it and betting their business on it. Well, they may not have been 10 years ago. And they understand how to deal with that from a legal standpoint. Yep, the legal requirements have been pushed through. The idea of a subscription model is not foreign to companies anymore and they like that model because of what it does at CapEx versus OpEx and where they can start using their budgets. That's actually helped now to drive and foster a market where you can build a successful open source company very well. And I'm not sure you could have a decade ago. It was a different market condition. Well, I mean, but Red Hat was built over a decade ago. Absolutely. All right, so that was sort of the exception that proves the... Well, the market was many confused at the time in OS, it's like sparks and so on and so forth. But the question I wanna ask on the scale side, so on the scale side it's pretty simple. You can just have people, costs, and bookings. And there anything else that you look at on the scale equation for you guys in your business as you throttle up? Well, it's people, costs, and bookings, but the other thing is the key partnerships and making sure those partners are enabled to go drive Hadoop as part of their platform they take to market. So whether you're Microsoft or Teradata or SAP or HP, which we just announced, or any of those, right? All of those companies have their platform that they can go drive Hadoop through the market with their platform. And those are all companies we work very closely with and that helps drive sales. Well, let's drill down on the double click on that because the big conversation has been enterprise ready. Is there enterprise ready? And everyone's like, yeah, it's people, enterprises are using it. So whether it's Cloudera, Hortonworks, you guys have a lot of success in there. We've heard from both companies. Some successes have some good scale and customers in the enterprise. The question for you is, talk about the partnerships that you've announced here that you're on top of the data platform and what are the key drivers around these new guys coming? Because SAP is big, they have HANA. We're going to hear from them earlier. That's also in memory. It's got a little big data. That's not open source, but supports Hadoop. So how is this all fitting in? So PurePlay Hadoop is by standalone is great. Now you've got existing environments like SAP and your partnerships. How does that work? How do you see the ecosystem evolving? So I was taught this analogy very early on, fish where the fish are. So if you want to think about it that way, if some of these large partners have great relationships with their customers and companies have better architecture on them and work with them, much easier to say, how do you fit within that world and make Hadoop consumable through that lens for that company in the way that they want to consume it? So let's make Hadoop consumable the way they want to consume it through the partners they want to. And if they want to do it independently through us directly, fantastic. They want to do this through SAP or through Teradata or through HP or through Microsoft. Fantastic. We support them all and not only support them, make sure that Hadoop makes their platform stronger and allows it to be more complex. Your business model is agnostic on the delivery indirect or direct. You still got a subscription. It's to a customer, whether it's sell-thrown through SAP, doesn't matter to you guys. It's still revenue, right? It's with the customer and the trust. If they have trust with a certain company, you don't care. Absolutely not. And just to give you a point, I just walked out of here in a meeting with the CIO just downtown. As we went through that, we're in the meeting for an hour going through this first time meeting with them. I walk out. Who's walking in right behind us? Our partner, Microsoft, is walking right in behind them to have this similar conversation. So it's clear that those relationships do matter. And each of these companies have trusted relationships in the environment. Talk about the HP deal a little bit more. So Vertica is like this hidden gem. Inside of HP, and they're doing a lot of good business. We did the Vertica user conference. It was very impressive in terms of what's going on there. I say HP's so big, it kind of gets lost. It's like the government sometimes. But so what's going on with HP? I would have expected they would have announced a partnership with you guys a long time ago, but it took a while. But what was announced and what's your relationship with them? So I think it's here. So what was announced is HP is a partner of ours. We work closely on how we're going to make Hadoop work with their platform. And also they can be a reseller and distributor of Hortonworks. So that part is in place. HP has fantastic relationships with their customers and contractual and buying vehicles, with trusted relationships, so they want to work with their customers. In terms of the opportunity of where it can go, is you look at where HP fits within that environment and how companies look and say, what else can I do with HP on the technology side? What else can I do with them on the Vertica side? And again, how do I make all this work together seamlessly in my environment? Yeah, so obviously it's not just Vertica. I mentioned Vertica, but HP's got autonomy. They got all kinds of different assets there. Absolutely. And so it's obviously a channel for you guys. It's great. You guys have been building up your channel. You're using the playbook that you know so well. But is there co-development also going on with HP? So absolutely. With HP, we're also looking at areas on co-development in terms of what we're going to do. Not all of those are set or announced yet, but that is a big part of it. And that's a great point because all of our partners think the co-development is what makes the difference between just having a partner and having a logo on a website and saying, are we actually improving the platform for the overall good of the community as well as the platform to help their products work better with Hadoop? And it's both. And that does require co-development. And that open source model is the right way to make sure all of that flows back into the same horizontal platform that everyone uses. Okay, so talk about the data platform you guys announced. Pre-announced prior to the show. What's the traction like? What's been the feedback? You've been on customer visits. You mentioned CIO. What's been some of the conversations and traction you've had with that? So the announcement was around Hadoop 2.0. And we were super excited about Hadoop 2.0 because this has been a vision for four years from many of our key founders, which is Hadoop has been phenomenal as a batch analytics platform in terms of what it can do. But what we find is for Hadoop to take its rightful role in the data architecture and effectively become a data operating system, become a data platform that companies can use, they want to say, how do I have the data in one place form a data lake or put all my data and land it in the same place but how do I interact with it in multiple ways? And before 2.0, you didn't have that capability. So 2.0 opens up to Hadoop to say, if you want to do batch analytics, fantastic. You want to stream data in through storm, fantastic. You want to do search against the platform. You want to do interactive query and real time speeds, three, four, five seconds. Absolutely, it's all there. You can do all that in the platform today and that the right way to do it is create Hadoop to be that enterprise viable platform. But now on top of it through Yarn, which is the resource management layer, open it up that you can unleash the waves of innovation of the rest of the community to go build on top. So you're not saying we're going to do it all. We're saying, let's open up innovation almost at the app store and allow that to flourish and everyone else to go build to that common platform. Well, we've been talking. We need more apps waiting and starting to see them come out. You're starting to see commercial off the shelf applications, which means that way more companies can get involved in this, right? Oh, we're in the early stages of the Hadoop market. We're focused on the platform side, but what will happen it is unleashing waves of innovation around the other technologies that move around it, applications that could go consume that data. Absolutely. Yeah, I mean the internet guys, the Wall Street guys, they have the resources to build the apps, but the mid-sized companies don't and that's the engine of a lot of the economic growth. Mm-hmm. But now think of again, partnership like Microsoft. Think of the ISV community that Microsoft has in terms of them and they're all building apps now to go leverage that platform to go to market for that community. Herb, talk about the customer center. Talk about how customers migrate to the new platform and you guys have any tools that happen. Mm-hmm. I mean, obviously, sensibility is key to innovate. What's the situation with customers? What do they need to do? It's a great question, because what we see with customers is the typical adoption pattern starts off with some level of a pilot, where they wanna try the use case, whether it's for funding reasons, technology reasons, how does it fit in my architecture? They're gonna do a pilot and they may do that pilot with us or with anybody else, but at some point they get through that pilot and say, this has real applicability. I can really get value of this. Now I wanna go broad. I wanna go to something more strategic. I wanna build a data lake. I wanna go broader. How do I go do that? When companies get serious about it, that's typically when they're coming to us, as they've gone through that stage and they wanna come to us. And when we have that conversation, we realize we need to make it easy to move from whatever other distribution they're on, whether it be an earlier version of Hortonworks to whether it's Cloderas, Mapours, anyone else's, need to make that seamless and dead simple for them to migrate. So, based on all the experience of the tens of migrations that we've done already, we've packaged this up into our migration kit that our customers can use. You have a turnkey migration kit. Turnkey migration. We can go migrate and do the estimation, automated migration and testing of that platform from another distribution or an early one of ours very seamlessly. Great. Okay, we're getting tight on time. Last question. Congratulations on all the success of the show. Being kind of a master operator and also experienced in the open source world, share with the folks out there your vision of how you see the world evolving over the next year or so. Obviously, things are firming up. The ground is validated around Hadoop. It's a market. There's demand. Things are firming up. There's some ground that needs to be hard enough a little bit, but the business is growing, this overall is growing. What's your take on, how things are gonna play out? Here's how I would see it playing out. 2013 in many cases, 2013 was the year of the pilot and the commercial enterprises. How do I go pilot this and start to get visibility on what Hadoop can do for me? 2014 is the year that many of those pilots are going to broad production and they wanna go deep and wide and the adoption cycle of how quickly they grow with Hadoop is amazing and it's because of the volume of data that they wanna be able to track and manage. So I see all those pilots go into commercial production and go in much broader and then the next tier of companies coming in and starting their pilots in 2014. And what I do think is that will start to unleash that next wave of innovation as we just talked about with applications and other things that can take advantage of the core platform, specific by vertical, specific by domain expertise and other things. Multi-vendor, all of the hardware, cloud enabled, all that stuff. And I'm confident that the partners we have and the others are the right companies to go advance that to the next state and take this market forward. Just one more little final point while you're here, it's a great perspective. Talk about the open source dynamic. I mean, you got open stack, looking back at the history of open source, I'd say we're in our fourth generation. I mean, I didn't talk about it. I mean, I'm in late 40s, so like we've seen the early days, but it's such a force, as Dave mentioned, of innovation. Has there been any updates to the modern era of open source maturity? Have you seen some new dynamics or is it the same game as always? So I think here's the dynamic is as open source has matured and I think I've come to realize open source, it's not just development. Open source is a development model and a business model. On the one hand, it's how do you do development and capture the thoughts of a community and all the expertise and can you go out innovate any individual company? And second, what's the right business model of how are you gonna go put that out in the market? Lower the barriers of entry, increase the ease of adoption and make it seamless for everyone to start using it. And now that companies have seen that, I think you'll start to see even more ways of innovation around open source. And you'll see, as you see, it's standardizing them. You see open stack, you're seeing people competing against essentially one big incumbent. So no one player can outweigh a community. That's essentially the same model, right? Same game. Essentially a community can out innovate an individual company. You've seen that with Hadoop and some of the early advances and other platforms already Hadoop that's all built within the core today. Great insight, President of Hortonworks. Great to be on theCUBE. And again, thank you for your support. The community, we appreciate it. We love open source. We're all about open source content. Again, we were here four years ago when I've never heard of Hadoop. We've been to all the Hadoop summits. Thanks to you guys. You guys do a great show Hadoop summits coming up. Is it Europe's coming up next for you? So we've got Europe in late March in Amsterdam and then we've got San Jose in June. Great, we'll be there at theCUBE. This is theCUBE live in New York City for big data NYC covering all the big data business action Hadoop world across the street in Strata Conference. We'll be right back after this short break. Thank you.