 provide new gadgets and talk about the latest in gaming, but those shows are just the tip of the iceberg and we're here for the tween technology consumers and those who live in the business today. And our viewers are getting begs for a program to fill that up. We're not just talking about how to get off the path that provides you with the creep, but we also want to analyze the big question that no one else is going to ask. Our guests aren't just here to provide common good. We work with people who know the industry from the inside of the app. The tech business isn't treated as a surface expansion of covering new equipment or prizes from start-up to every new movement in between. To think of us as a companion to the community. We're here every morning trying to extract the signal from the noise. We're the cubic cells in event coverage. We're working to bring that experience to you consistently every morning. We use the top stories of the day to provide you with breaking and analysis systems so that you can forecast future trends. We're here before you even wake up. We're creating a fundamental change in laying the foundation setting the standard. And this is just the beginning. Going to Disneyland. Going to Disneyland. I mean these guys are great. I think this is a revolutionary forum. Up till a few years ago I'd never seen this in my entire career. These guys are great interviewers. They're spot on. They're sharp. They're funny to work with. And they just ask great questions so it's a real pleasure to be on the queue. It's really great. What's so neat about it is it's like real-time discussions and also just being able to have people share their views simultaneously. So I love it. I think it's really fun. It's a great way to get the message out and to have a dialogue. This is a fantastic way to have the conversation with guys who know what's going on. Who can kind of scratch below the surface. Who can respond to what's happening right now on a Twitter feed about maybe some technology that's been in the marketplace and respond and have a conversation. So it's a way to kind of demystify what's going on for a lot of folks. And for me as a marketing guy I'm also really keen on the huge community that has built up and follows every single post that you guys have got. So it's fantastic, irreplaceable. Actually it went really, really well. I mean the thing that I really like about theCUBE is you guys get, I mean bottom line is we can talk about high level strategy. We can talk about execution. We can talk about competitive and market. And what I like is the interactive banter back and forth. Plus the fact that when I think about some of the conversations we have they're not only deep. They're not only rich. But the audience themselves will really come to benefit from those conversations. Also I think theCUBE you guys always have very thoughtful questions. Really insightful comments and it actually makes for a really fun discussion. I want folks out there to understand the depth of technical inspection that goes on with you guys. It's deeper than most analysts we talk to. I mean so we roll up our sleeves. We'll spend a half and a half day on the hot new technology instead of the PowerPoint ivy that goes on a lot of the time in our industry. So it's, you know, it's when you get a perspective from theCUBE that is, you know, that's as good as a validation. Here live in Silicon Valley, we're at the heart of Silicon Valley. We're in San Jose, San Jose Convention Center for Hadoop Summit. This is the second day of live wall-to-wall coverage, Silicon Angles theCUBE. This is our flagship program where we go out to the events and extract a signal from the noise and we'll go where the action is. If there's a story that needs to be written, we'll do it. If we're going to go out to the floor and do an interview, we'll do that. But here at theCUBE we bring in folks who have that signal. Those tech athletes, CEOs, entrepreneurs, developers, and occasionally it's people who build huge production Hadoop systems from either Facebook and now working at Nutanus. That's our next guest here to discuss that. And we're going to have a great conversation around this new model, the new modern era of software, software-led infrastructure, software to find everything, and obviously Hadoop and Big Data. I'm John Furrier, the founder of Silicon Angle. I'm joined by my co-host. Hi everybody, this is Dave Vellante from wikibon.org. Karthik Ranganathan is here. He is a technical staff member at Nutanix. And Nutanix is a company that has really had a great deal of success in the hyperscale, bringing a lot of that success into the enterprise. Karthik, welcome to theCUBE. Thank you. Very excited to be here. Yeah, so we have been tracking Nutanix for a while. Everybody talks about converged infrastructure. You guys built converged infrastructure from the ground up. You came to Nutanix from Facebook. What attracted you to Nutanix? So at Facebook we used to manage, we do manage systems of very large and massive scale. So the focus is on performance and the focus is also on manageability and scalability, right? And what my vision was, let me do this for the industry at large. So that's the thing that attracted me to Nutanix. The vision kind of coincided. That's what Nutanix is doing. The convergence, the scale out. So that's the angle. We talked a lot on theCUBE. John and I about the differences between the hyperscale, the pure play hyperscale markets and the enterprise markets. And hyperscale, of course, is well documented. A lot of really smart people, PhDs, engineers running around. They will spend time on a problem to save money, whereas the enterprise, they don't have as many resources. They have a lot of resources, but maybe not as deep technical PhDs. They'll spend money to save time. They'll buy a box. That's right. They're looking for companies like Nutanix to provide that capability. And we see the hyperscale mentality seeping into the enterprise. So is that essentially what Nutanix is trying to do? Is bridge that gap and talk about that a little bit. Exactly. I think you hit the nail on the there. So with Nutanix, the idea is it's a device. It encompasses all your performance needs by putting in flash, by putting in RAM, CPU and disk in the ratio you need it, so that you can configure a distributed cluster to get to achieve your end. And it's all converged. So the nice part is each of these devices has a distributed system built into it, a file system. And you can just add more and more of these and scale out and grow as you need. Now the hard part was when you live this in your world with Facebook, when you scale out like that and you're using a shared nothing environment, sometimes things get a little lumpy or rebalance things is sometimes a challenge. So companies like Facebook, you write a lot of software to do that. How will the enterprise deal with that problem? You're saying Nutanix has that magic sauce already. You're building that out. Where are we in terms of the maturity of that software for the enterprise? Sure. So let me break it down in a slightly different way. The magic sauce is a continuum over time. So you see more edge cases, you fix them. You see different kinds of use cases, you adapt. You see more and more data, you scale up to it. So I would say Nutanix has a lot of the magic sauce and we're continually working to put more and more of the sauce with time to stay relevant, right? And that's the name of the game. That's the fun part. One of the things we talk about theCUBE and we look at the landscape of the big data world, you've got to see the pure Hadoop plays. Open source has been a big part of that. And now you've got the commercial vendors coming in. And Dave was kind of referring to the old way, which is to scale up buy a bunch of gear, buy an Oracle license with a database, boatload of storage, and use commercial software. Right. You lived and pioneered with Facebook. I was at your talk at HBase 2012. You gave keynote talking about how you rolled out Hadoop into production and all the nuances and all the benefits that it did. That's a new concept. That is a lot of people are looking at what Facebook has done with their ops and their dev, their dev ops, as the new way, the modern way to do engineering and to deploy quickly and scale. So in the world now, the enterprise grade as being talked about here, open source has led that charge. What do you see for folks out there that are in positions, that are on technical staffs, in large enterprises? What's their mindset? What should their mindset be? How should they look at some of the tooling and the platforms available? What advice would you give them looking at you now? You went from Facebook, now you're at New Times, which is pioneering kind of a new way. What is the mindset and then what should they be doing? So I'll say that the most important thing I feel is understanding your own scale. If you're going to build something that has to scale, then scale has to be at the forefront of everything. I feel that performance itself is a second, although a very close second to scale. So the architecture itself from ground up has to have scale in mind. So I think that is one thing, whether it's open source or commercial, it doesn't matter, but as long as the fundamental architecture would scale and would work with commodity hardware and would handle failures properly, I think that's the first step, that's the first battle one. And then immediately after that, you have manageability. You need to know if something goes wrong, where does it go wrong? How do I fix it? How do I troubleshoot my system? Will my system auto-fix itself for the common failure cases? I'll also throw in the other extreme of over-optimizing and over-automating where automatic fixes itself can lead to a little bit of unwanted behavior. So it's kind of a trade-off between the two, and I think that's what people should go for the most. Performance is always ongoing, so try to keep doing the performance stuff, but upfront built for scale and built for manageability. Yeah, there's an interesting point you're making about automation, because again, the hyper-scale is highly automated and whereas the enterprise, of course, is very labor intensive, it's sort of a culture shock for the traditional enterprise to trust the machines, to fix themselves, and there's a real skepticism there. So are you seeing your customers move to that balance, or are they resisting? Do they believe when you say you can come in and help them find that balance? So I think initially, like you said, it is a shock, but slowly working through the mindset and trying to talk to them about what sort of scale will they be seeing? What are the benefits of having the convergence? What is the ease that it can bring to their actual manageability story? So people usually turn and that usually hits a chord with them, so then they start coming around and try and play with the software and try to understand, and then there are questions about what happens if this goes down, what happens if that fails, what happens if on this type of a failure, and working through all of that, I think people usually see the benefits of a scale out and distributed architecture. So Karthik, what about Hadoop? We're here at the Hadoop Summit obviously, you guys are certified now on Hortonworks platform with regard to I think the first converged infrastructure you announced. So why converged infrastructure for Hadoop? What's, you know, the compare and contrast with bare metal and give us some insight there. Sure. So let me say this in the form of a story, right? Let's take a typical enterprise out there. They are either a, trying to figure out their big data needs and trying to see how they can put it to good use, or b, have a bunch of data in the few terabytes, maybe tens or hundreds of terabytes, and trying to get some meaningful information from it, not necessarily build elaborate pipelines. So this is a typical enterprise. So the story here is, would you rather go and try to figure out what sort of hardware you need, try to provision that hardware, try to figure out how to maintain it, try to get a team to maintain it, get a support contract, do your upgrades. That's a laborious process. So even if, like, and the traditional no for putting, for doing virtualization with Hadoop is the performance angle. That's usually what people say. Firstly, I believe that with convergence, you don't give up much performance because all of your data is local. But even that aside, the bigger story for the enterprises is the manageability aspect. You can manage, like, the people, the data centers right now are very, very familiar with virtualization. So you can run Hadoop in a virtual machine, and your admins will already know how to manage it. So that's the first part. The second part is, you would already have a virtualized story. You have VDI, you have server virtualization. You already have some workloads running. You can run Hadoop alongside that and put it to good use and see the benefits before deciding if you want a dedicated cluster. You want to go bare metal. You want to go virtualized. You can get into the depths of it as it starts proving it out to you, as opposed to putting an upfront cost and not knowing where it'll go. Talk about manageability with Hadoop. Obviously, you know, the age base has been very popular, but the criticism of age base is obviously bare metal is a key part of it. Supporting that and having administrators on it and managing age base. And there's some, you know, obviously the products evolving. Talk about your view technically about where the platform is right now with Hadoop. Obviously, the growth is great. Everyone's happy. Everyone's kind of rowing in the boats in the same direction. Maybe a few bumps and bruises along the way between different folks. But for the most part, the ecosystem is growing very rapidly. What needs to get done in the Hadoop ecosystem from your perspective to take it the next level up? So the first part is, I'll say that on the manageability side, I think it's not just the UI which tells you about how many nodes are alive or what is the utilization of your cluster. There are a lot of soft points. You would like to know what caused those failures. That analysis is usually pretty complicated. There is no one view that gives you the insight into the system. So obviously you're talking about age base. So from the past, let me talk about how we did it at Facebook. We used to have a dashboard which would summarize all our clusters. We have a lot of clusters, a lot of machines. At that scale you don't ever think, let me go and figure out what's wrong with these couple of machines. Your fingers will probably fall off before you get to all the machines. That's the scale. So you look at these charts and you try to zero down on what was going on at the time. Is it a network incident? Is it a disk incident? Are these a bunch of machines? Is it one cluster? Try to zero in as close as you can get and then you start jumping into the machine and looking at it. So that aspect is, like each of the customers have to do that for themselves. So that is one aspect. And so on. So we got to wrap up here. I want to ask you one more question. We have been very impressed with Neutonix as a company. They're here in San Jose area, Silicon Valley. Great founder, great CEO, great investors. But they're doing things differently. You're at Facebook, you're doing all this killer, awesome work. Why Neutonix? What are they doing that's got you really engaged to go join the technical team over there? What about Neutonix really got you excited? They're bringing all the principles that I was really passionate about at Facebook to the industry at large. That's it in a nutshell. So all of the details the distributed systems behind it, the commodity software, the failure resilience, the management story, the scale out, all of that stuff is all in a simple box, in a simple form factor that anybody can consume. So that's the best part. You guys built it from scratch at Facebook and now they've commercialized essentially the box. Again, Neutonix is a very impressive company. Thanks for coming on theCUBE and sharing your Facebook experiences. And certainly we'd love to have you on again. We'll talk to the folks at Neutonix. We'll convince them to let you be a Cube regular on our news program. But ultimately that's the real deal. How do you commercialize what the pioneers built from scratch? And I think Neutonix's performance as a company will dictate the proof points. Congratulations. This is theCUBE. We'll be right back with our next guest after this short break. Again, this market's growing. It's changing. It's exciting. Hadoop is really, really growing and lifting. And we're here to bring it all to you from the Cube. SiliconANGLE and Wikibon's coverage. We'll be right back after this short break. We looked at all the programs out there and identified a gap in tech news coverage. There are plenty of tech shows that provide new gadgets and talk about the latest in gaming. But those shows are just the tip of the iceberg and we're here for the deep dive. There's a difference between technology consumers and those who live the business day today. And our viewers recognize that. The market begged for our program to fill that void. We're not just touting off headlines. Our goal is to provide you with a story but we also want to analyze the big picture and ask the questions that no one else is asking. Our guests aren't just here to provide commentary. We work with analysts who know the industry from the inside out. The tech business isn't new but many networks treated as if it is and really barely scratch the surface on technology coverage. We follow the expansion of the cloud and the evolution of big data. We're covering new enterprise from startup to IPO and every move in between. So what do you think was the source of this misinformation and so you mentioned briefly there are several other. If that's the case then why does the world need another software as a service player? I like to think of us as a companion to the Cube. We're here every morning trying to extract the signal from the noise. Where the Cube excels in event coverage we're working to bring that experience to you consistently every morning. We use the top stories of the day to provide you with breaking analysis so that you can forecast future trends. We're here before you even wake up. We're creating a fundamental change in news coverage. Laying the foundation and setting the standard. And this is just the beginning. Inside the Cube we are live in Silicon Valley. We're in San Jose Convention Center. This is siliconangle.com and wikibon.org's exclusive coverage of Hadoop Summit 2013. I'm John Furrier. This is 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. Join with my co-host. I'm Dave Vellante of wikibon.org. Welcome everybody. Ben Werther is here. A longtime Cube alum. He's the founder and CEO of Plattfora, a company that's bringing sub-second interactive into Hadoop. Ben, welcome back to the Cube. Thank you. No, I'm just saying, Cube alum that you've been on from the beginning of the Cube. And now you're a big time CEO, big financing, growing, big company. What's it like? Really, really exciting. I think we've, you know, I'm coming on when we were at the point where we had a concept. We had the beginnings of a product. And I think I joined you guys around the time we launched that product. Now we're in the market. We're seeing, you know, we're adding a couple of 4,500 customers. We're really getting momentum across different verticals. So Dave, Dave and I were talking the other day and we're talking about the changing nature of this modern era. And one of the things that's going on in the enterprise right now is you have the old school and the new school. And I think, you know, it's really clear to us that we've talked about this in the past, old way, new way. And I think at this show enterprise grade, Ben, has been key. And you guys take a different approach. So give us an update on what you guys are doing. And then let's talk about what you guys are doing differently and how are you guys enabling that innovation for the customers? Sure, absolutely. You know, I think the old way is, you know, the question is sort of how do you go rebuild the old data warehouse on this new stack of technology, which I think there are use cases where that makes sense. But in, you know, one of the things that sort of struck us more and more with the customers we work with is, hey, the kinds of data you're putting into Hadoop, and you want to mix together, these are not data sets that you want to treat the same old way. You know, you don't want to do traditional sort of state business reporting against, you know, these event logs and clickstream data and sort of social graph data. There's so much more interesting questions about behavior and insight you want to drive from those. And so I think it's really pushing towards not just how, you know, how do we make BI work in this environment, but you know, really how do we allow people to kind of express these things that today take custom development, they take a year of smart data scientists and make those questions that a business user can just sit down and ask visually, you know, this afternoon. Well, that's what I like having been on, John, because you know, let's face it, a lot of the platform vendors here, they're coexisting, they're working with other large, you know, database companies or traditional BI companies, and they've got to be politically correct. And you are actually quite kind just now. But I mean, you've been more forceful about, you know, look, yes, you can do that. But there's so much more business value that you can drive if you kind of rethink the way in which BI is done. And we've talked about this before. The BI in many ways has just failed to live up to its promises. So what are you seeing today and what gives you confidence that your vision will be able to live up to its promise? Yeah. I mean, I think the clearest indication is customers using the product. Nothing gets me more excited than going into an organization and there's often naysayers who say, hey, we have these existing investments. How do you leverage whatever, you know, the micro-strategy and the other tools I have. And our point is, okay, look, that's fine to have for the existing use cases. They're good products. But let's talk about the kinds of things you're trying to answer, the kind of data you're bringing together. And let's look at what you're trying to achieve. And we rapidly see these business users disenfranchise data sets that often have much more interesting character. If I'm getting clickstream data, every record may have dozens or hundreds of different tags of different things that I can use for interesting segmentation, different types of behavioral analytics. I can look at combining these data sets in surprising ways. And the idea that you're going to try to do that, the idea that even if you made the traditional stuff work in Hadoop land, which I think is a dead end, honestly, because, you know, when you get there you realize that that's not success. You just recreated a model that had a lot of inherent flaws in it. So when you go into an account, if you're talking about the light of business person, the last thing they're going to say that their objective is is to preserve their current data warehouse infrastructure. They don't know if they don't care about it. But there must be a segment of the population that says, okay, yes, we want all these great insights, but we also want to leverage this investment that we've made over the last 20 years. What do you do in that situation? Do you run, not walk? No, I think we're very clear with them. You're not going to throw away any of your existing investments and there are a lot of good use cases that are being served by them. But I think the biggest learning that companies have is they start out saying, okay, I have Hadoop and I'm going to... I have a data warehouse. I add Hadoop and I want to pre-process data in Hadoop and put the arrow goes from Hadoop into the data warehouse so that the real work happens in the data warehouse. And I think the challenge is you end up with silos of data. You end up with a model where Hadoop doesn't have all the data and the data warehouse doesn't have all the data and you have frustrated customers, frustrated users. And the next step of maturity is you flip the arrow and you start putting all that data into Hadoop so you still have your existing systems. But there's a whole class of new questions that you can enable because you now have data that may span much like a wider variety of silos. Some of the banking customers we work with, they're thinking about how do I do a 360 degree view of a customer analysis with all these different silos of data that if I was to try to integrate in a traditional sense, it would be a decade long project. But just landing that data in Hadoop can be the beginning of doing this very, very rapidly in an agile fashion. Well, the problem with that, what you described is you get business processes established for each one of those data silos that are very rigid. So talk about how your customers are changing their business processes. One is we sidestepped that a little bit because we don't try to replace the existing systems. They're doing transactional work and it's working in those environments. That's good. We don't want to be in that business. But I think that the key is that data is a resource that can be combined with other things in a way that we're not trying to go in and replace the existing data from your reports. We're trying to help you answer more interesting questions than aren't being answered today. So the bar is very, very low. And so once you get those data as long as you can start to unify some of the common, you know, if I'm thinking about 360 degree of a customer what is a customer? Do I have keys that I can match up in some way? Can I view this in a holistic way? But then I can, if I add new data being generated, if I get into mobile advertising and I have all these different new events related to mobile behavior, as long as I can key that to customer I can weave that in and have this sort of dimensional, entity-centric view of it rather than the traditional style schema which becomes like this heavy, heavy-weight thing to manipulate. And I want to ask you about, so I was in the news here you had some certification. You guys are certified with Hortonworks now. I see you certified Technology Party with Cloudera as well. You're in the ecosystem, right?