 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 for Silicon Valley Big Data. Big Data SV, as we call it, is an extension to our Big Data NYC event a few months ago in New York City, where theCUBE goes and extracts a signal from the noise, Silicon Angle and Wikibon's production. I'm John Furrier, the founder of Silicon Angle. And we have two CUBE alumnus here, Bruno and Richie, here to debate open source versus proprietary, but it's really not kind of a debate because Bruno is a fill-in for Eli Collins and Cloudera. But we just want to have a general chat because I know you can talk on both sides. You've been on that. So guys, welcome back to theCUBE. But before we get started, Bruno, we haven't had a chance to talk. We saw each other briefly last night, towards the end of the day. You were at the startup competition at Strada. You gave away a Vespa and you're going to be coming to our party tonight. So just before we get into the debate, how's the show going for you and what's new with you? Yeah, so the show has been great to us. We, as you know, we announced a collaborative data science. So we're bringing a new way for companies to harness both the power of their data and then leach to collab the innovation of their people. So that's what we announced. We had a great party last night. With Alpine Data Labs. Alpine Data Labs. Yep. And so we'll be talking more about that, you know, for the show. But in general, I think it's been good. The show floor looks pretty good. We have partners in Pivotal and other companies that we're working with. There's a lot of interesting conversations. I think the thing we haven't broken through just yet, and we talked about last time, is, you know, the analytics applications are, that's the killer thing we should be focused on. So, you know, I'm waiting to see tomorrow what we're going to be talking about on that area. Rishi, welcome back. We just talked earlier today. We called you the professor because of your knowledge. And Professor of Big Data, CEO of InfoObjects, is doing a lot of work with customers. Guys, it's a really build out year for us. And some people are saying, hey, it doesn't matter what religion you're from, open source or proprietary, a lot of build out. But still the debate continues. So there's a dynamic going on where, you know, if it makes sense to be proprietary and you can harden that top and extract the way kind of under the hood, maybe that's a good thing. Functionality could be the new open source. So, but yet pure open source is a trend. So Rishi, we'll start with you the philosophy of open source versus the for-profit enterprise focus which is lock-in. And at the end of the day, the old school was lock-in. But with open source, the dynamics are changing. What's your take on that? Yes. So on one side, you used to have the old school proprietary in which there is a complete IP of the vendor. But what's happening now, what they're calling it open core. So you have the open source and then you're providing proprietary above. So that is still vendor lock-in involved there. And what's our take is that all the big data problems you can solve using pure open source software. So, and what's happening is if you see a lot of proprietary vendors, the kind of things they had for Hadoop. For example, snapshots or the high availability or penetration and all, that's all available now. So you build your, even your open core approach, you add two more features, right? And you charge customers for that feature. But after six months, if you have a very good stable version of that feature in open source, how do you compete? So the only place where I think you can compete is the customer support, customer satisfaction, customer turnaround and things like that. So basically the whole professional services model, the whole services, services is the only place where you can make money. We are a pure services company. And as I was saying, if I was back, we brag about not having any IP, right? So that puts a wrinkle in, say, Cloudera. Or does it? I mean, you've got Hortonworks. They're going the red, kind of called the red hat model. But Cloudera is involved in open source. They have a lot of contributors on the core, Hadoop and a lot of the projects and Paula's open source. Are they open source? They have some proprietary or some... So yes, Cloudera is a great example. So what happened was that they have this Cloudera manager for which they have been charging money. And Hortonworks came up with Ambari, which is open source. So now how do you compete with it? Cloudera has a huge first more advantage. They've done a great job in contributing to the community. The same thing Hortonworks has done and a lot of other companies also. But the question remains that you come up with a feature which is proprietary or you call it part proprietary. But somebody else comes up with the same feature free of cost. How do you compete? So if I can confess my bias, I probably should give you some background before I go there because before working for Alpine and a few other companies in the valley, I worked for Microsoft. So I guess my depth in open source is probably not as deep as the professor. You're okay punches, come on. Let's think up for Microsoft. But I will say and what's interesting is you were asking about Alpine data labs. We actually, the product, the commercial product we just shipped is based on open source called OpenCore. So we have both these tracks where we have a support of the community through open source software and the advancement of the community because I believe in what you're saying. And we also have the advance of commercial solutions. I think the thing where I struggle a little bit with the argument being one versus the other only, I think actually the truth is somewhere in the middle. You know, if I, we have Jeffrey Moore that did the keynote, I think it was this morning, and he talked about- We seem to be crossing the chasm for like 10 years. Are we ever going to cross the chasm? And what happens when you do is- No, he doesn't get the speeches anymore. There's no openings for a speech if we never cross the chasm. And there's a bullying ally. This is very confusing analogy. But he does have one analogy that I think is very powerful that I read in these books a while back. It's the idea of the whole product. And the whole product is you have to have a core software. You have to have a set of services around it, which might be the business you're in. You have to have a community. So, you know, it's difficult to say, do you need software open source and software commercial on these two things can be. I think in the end, the combination of the two could be the answer. Well, that was the argument which was, which had been working from last few years. But as you see more and more stable open source products coming out with all the features which proprietary has. So my question is, what can you provide in proprietary when everything which a customer needs is provided in a stable open source Apache product? Well, can you pick an example? For instance, I mean- I just gave an example of- A specific example of the product. Cloud Data Manager versus, again, I'm not supporting hard-on-works or anything. But if you see that they made body open source, right? And so what happens? So- Now Cloud Data is making the Impala open source. All right, let me throw a wriggle in this. So, a question for you guys is, how does a company compete in an open source framework? Because, you know, one strategy might be, hey, you know what, I'm going to just hire all the best dudes in open source and I'm going to own the project. Yeah. Just in the old days with W3C and ITF, the standard bodies, you would slow it down and the proprietary technology would come in and end up slowing down the open movement and then the proprietary technology would take the ball and run with it and score the touchdown. So here it opens just a little bit different. I mean, Cloud Era does contribute a lot of core people to the project. Does that mean they have a de facto presence? So what is free? Is it really free? Software? Well, so, and Cloud Era, as I said, has a huge first moral advantage and we all should be grateful to them because of the number of contributors they have provided to open source. Hortonworks has also taken a similar approach, but whatever small proprietary piece which Cloud Era had, even that Hortonworks has removed. What they say is, if you call it proprietary, it's just they are bundling, which they call it HTTP. So the only thing they are doing is bundling, which is not even proper. They are calling it not even that proprietary. They are saying even that is open source because you know how that bundling is happening. And then what they are doing is that they are trying to focus on the pure professional services model. So, and my point is everything has to go that way because what you're calling is proprietary is some extra feature, some extra addition. And somebody in a month or two months or a one year time will come up with an open source alternative of that. And when there is something free available and which is also good, it's impossible to get people to pay money for. So if I can be a little controversial, maybe disagree with you, because I do think that this idea of one versus the other or the term everything that needs to go one way, I don't think that's the reality. I mean, we struggle with this ourselves, like I explained earlier, I'll give you a specific example. We have OpenCourse, which was an asset that Pivotal had and that now open data labs is a custodian force that we contribute to the community and so forth. But at the same time, we're using this code into a commercial product and evolving it for scenarios that require support, specific functionalities that customers in particular industries are going to need, that maybe the community doesn't care as much about developing. If we're focusing on financial services, you get security requirements there that are required, and this is not service, this is a hardcore software development. The question is, and I think, John, when you're asking this a little bit earlier, is that what is the business model, right? If you're in the open source and you're betting everything there, then your business model is services, sounds like. If you're on the commercial side as a software company, then you have to orient your business model towards the development of software and the support of that software, first and foremost. No? Maybe you've described that. So you took it into the right direction, and the direction is this, that you have the open source software, and then you would have the vendor, sorry, not vendor, but the vertical specific solutions, right? So that's what you would develop, and if you are doing it for 10 customers, one after another, you would always have some knowledge-based develop which you are going to reuse, but that would still stick to the solution space rather than the software space of proprietary software. So I would call it more of an application space, so you would develop applications for the client's needs, and that's what we say. We say that any specific thing which a client needs and every client would need a lot of customization, so we would do it for you, based on what you want, what your need is, and you own the IP, the customer owns the IP rather than we own the IP. I think the open source is pretty much an obvious home run. People love open source as just a gift from the heavens when you think about it from a computer science standpoint, so we were all there. I think we all agree. The question is, how does someone have a competitive advantage, a startup? Is it, I mean, the old days it was dual source licenses, the end of different licenses, you had the slice and dice. Now, if you go open core and you say, hey, let's use the contribution model of projects in a way it's a democratization of IP, but yet companies still need to have a lever. You're saying support Rishi, that's a key differentiator. Is there anything else? Yeah, software, software. I mean, what we see, we have this debate right now. Like I said, open course is an open source collaboration platform. We have taken this code into Alpine's commercial software that adds events analytics and collaboration to it. On one end, we have a very deliberate approach to the market. We say we go after these verticals, these types of companies and we build for them. That's kind of like the different more tactic, which is you build around the set of set customers with their needs. However, on the open source side, we uncover a lot of innovation that the community will build. Like our plug-ins, today I had a customer called Oridia that started building using the open source code and built R into that, which technically is the stuff that we could be doing. We just haven't been looking at that. Now, the way they're differentiating is not because of providing services, they're actually building software based on the open source platform that has been available to them. So I think saying that service is the only place where you can differentiate in open source. I don't think that's the only, I mean, it might be the opportunity for many companies. So you're saying you don't have to contribute the code back so there's a balance? In our case, in the case of Alpine, we contribute back. Oh, you guys, you're saying as a philosophy, an opportunity for an entrepreneur to be like, okay, I'm gonna patch some projects, I'm gonna use this open source and as part of my extracting value of giving back to the spirit of open source, they'll contribute code back to open source as a quid pro quo, but yet do something unique that they'll hold on their own. You'd have some features that I think, sometimes as a company, you have to make some strategic choice and saying, I'm gonna pick a stupid example, but just for the argument, we're only gonna build for the Asian market. I don't know, it's a bad one. And so you go out as a company, as a commercial venture, you say for the Asian market, we're building this. And then somebody in Germany says, you know what, I'm building this feature for the German market with these set of criteria that are unique to me. Data compliance. And then all of a sudden, for instance, yeah. And then all of a sudden, as the company that might be the leading developer in this market, I might not have focused on the German market, but now I'm realizing, wow, this community here is actually bigger than I thought it was. And now it's taking my attention as like, you know what, I wouldn't have spent money on development of that, but I realized this community was a total blind spot for me. And then I have built this, and that's the way they differentiate. That's software, that's not service. You have taken it to a completely different dimension, which is like, because different geographies have different markets. But you can say in the streets, you could pick, you know, if you focus on retail. Specific solutions is a different dimension, I would say, because if you develop, say for example, healthcare specific solution, now would you open source it? No, because that's kind of the knowledge base which remains with you as a consulting company or as a consulting arm of a company. Your tooling becomes an advantage. So you can take tooling in a direction. You go platform, open core, contribute and have a presence there and you win or lose by how good you are. So you're standing in the community is also marketing. I mean, in a way that's marketing, right? Yeah, absolutely. And I go back to the definition of Geoffrey Moore. What's the whole product? It starts with the software base. It's wrapped around a community of partners. It's wrapped around a set of things. And I think at each layer, you can compete better. You can add to the software stack. You can add to the service delivery. You can add to support. You can add to the go-to-market. Okay, I gotta ask you the Geoffrey Moore joke because I mean, this is two years ago, he said we were crossing the Casemont Hadoop Summit. We interviewed him in the queue. He's a Cube alumni as well. By the way, the two Microsoft board members are Cube alumni, John Thompson, chairman of the board and the new CEO, just thought I'd tell you that. So we got a Cube alumni there. And so, but that's two years ago. What happened two years ago? Why haven't we crossed the Casemont? For this base? Yeah. Depends how big the Casemont is, I guess. That's it. You gotta take it. I think there's a lot of innovation, but I think what's happening, at least for my space, so I can't talk about everybody, but it takes a while. If you look at the evolution of the enterprise data warehouse, it took us really 10 years maybe, probably even 15 years to take full advantage of the apps built on top of it. And so I think we're kind of reliving through it. And to some extent, I think we're making the same mistakes we made in the EDW world. We assumed that this thing was just for a subset of the population. We didn't go to democratization faster than take, well, and so I think we're paying a little bit for that tax. We're definitely faster. I don't think it's gonna take 10 years, but I think it's reasonable to say two, three years. It's probably where we are at now. So, enterprise data warehouse market, I think it's a good comparison, but at that time open source was not a factor. Okay. Don't you think that makes it a little bit different? So you think having open source now is gonna be an accelerator? It's going to make it different. For example, I talked about Cloud Data Manager versus Ambari. Now let's take example of Datamere. What Datamere has done is they have a proprietary platform for visualization and analytics. Don't say anything bad about Stefan, okay? He's a good friend of mine. I'm just kidding. No, no, no, I love the platform. No, no, no, he's super smart. I love the platform, but we as a consulting company, for example, we are doing a lot of work in D3.js. What if somebody tomorrow comes up with an open source platform in D3.js? Then they win, so I don't know how do you how do you keep selling this kind of proprietary stuff? And you're doing, I mean, as I said, Cloud Data did a great job. Datamere is doing an awesome job. And they also have a huge first move advantage. I think it comes down to this, and this will tee up the last question for you guys, because that's a good point. The value proposition has to define the price. So if it makes more sense for a data mirror or a company to have a unique proprietary solution, if the value's there and there's a market, great. Now the beauty of open sources, you can lower costs and deliver the same functionality. If you look at what's open source has done, it's been one of the most leveling successes disrupting. You can just say, okay, I'll put these projects together and the incumbent is charging this for that proprietary and you come in and disrupt the market by a huge order of magnitude. So the question is, that's a well proven formula and a lot of people have done well with that. So the question for you guys is, where we are now in open source, what's the next generation of open source gonna look like? Okay, if you think, okay, I remember back in the days in the 80s, in the 90s we're standing on each other's shoulders of giants, what is the maturization of open source going to look like the next generation? What is it going to evolve into? Professor? Okay, so I think more and more advanced software is going to be contributed back to open source. For example, Cloudera developed Impala and now they are putting it into open source, right? So that's what's going to happen, more and more advanced software and more stable product is going to be in the open source. One more thing I want to add here is that if everything is going to be open source, how companies are going to make money? So that's where I would use analogy of get. So in the get, there is this concept of network of trust, right? So what's going to happen is, because Cloudera, because of their presence, because of their name, the market trusts them more than say info objects, right? So obviously they would be able to get more market, more customers, more traction than other companies. I think it would get down to that at the end of the day that how much market trusts you. So I'd say there's probably two things, one of which I think I agree with you. I think the maturity of commercial ventures in contributing more to the open source community, I think that's definitely happening. If you look 10 years ago, it was really a paradox, right? You either were commercial and open source, you got lots of companies in Microsoft, is involved in open source. Alpine is a much smaller company, of course, but also contributing back to the open source. So I think there's a maturity for commercial ventures to say, there's great value in the open source. I also think on the open source side, because it's now becoming easier to connect developers across the world, and we've got so many examples now. I think you're going to see successes happen very fast, but also you're going to see things dying faster. Because if they don't pick up as fast as they should, they will just kind of drift away. And I think in a way we didn't have that as much as we used to, right? You might have some projects open source, projects that would stay forever, and that wouldn't really contribute any value to the ecosystem. Yeah, but that's where Apache Foundation already has a very well laid out, very well established process that... But still, you find a lot of projects that have 10 contributors, right? But they won't become the full project that would be either incubator or at that status level. At what point do you decide, okay, this is a failure or it needs to be rebooted? Or how do you manage that? I don't think if you manage that. Yes, I don't know exact details, but I think Apache community, they already have this criteria that there should be so many contributors from so many companies, only then it becomes a real Apache project. Gentlemen, tech athletes, thank you for sharing your perspective on theCUBE. It's really great. Let's keep the lively conversation, because we're going to do a wrap up of day two, and of course we have the CUBE party here at the Hilton in Santa Clara, right across the hall. So if you're watching this or in the area, drop me a note or Dave or swing by the Hilton and join our party from six to nine, and who knows how late it's going to go. I might have to get a room if I have too many cocktails, but we'll see. This is theCUBE, lively conversation, open source is a winning strategy, but at the end of the day, it's software. I like that network of trust. I think the Git model is the way I see that too. I think it's going to be competition amongst foundations. I'm going to start open source competition, and competition is a good thing. Free software, innovation, that's the heart of it. Thanks for the conversation. Really appreciate it. We'll be right back with day two wrap up after this short break.