 Okay, we're back live here, winding down day three of three days of nonstop coverage here on Silicon Angles, theCUBE, our flagship program, we've got the events and extract the signal from the noise. I'm John Furrier, the founder of SiliconAngle.com and my co-host today is Jeff Kelly for this segment and we're here with entrepreneur CEO Ben Werther of Plattfora, launched startup, doing very, very well. You guys launched, Ben, welcome back to theCUBE. Thank you, thank you, always good to be back. We always love having you on because one, two, multiple reasons, you're an entrepreneur, CEO, venture-back startup, doing some really amazing work and really stole the show at Strata, Hadoop World in the fall with your positioning and your launch of your company, which essentially was redefining BI and then how that's getting done at a scale and performance no one's ever seen before or heard of or seen before. And the website was fantastic, we're all impressed with the design of the new website, but in all seriousness, you guys have really been working hard on your positioning, it's unique, it's about a forward-looking view and the talk of this show, as many themes are going on, obviously, Strata's got a lot of range, but the big news was the big commercial jockeying and competitiveness around EMC and Intel coming on the distro side. So that news aside, really it's about the BI market, the pre-existing BI market, data warehousing, put some SQL, Hadoop put some SQL, and target those existing folks that you were talking about that are sitting there like looking for a solution and need help. So the BI market is trying to be big data ready, so can you share your commentary on one that news and then we'll get back into some of the things you're doing, but I want to get your perspective on the recent news around this BI position in EMC. Yeah, absolutely, there's a few pieces to the story that everybody's been telling and a lot of it comes down to subtle SQL interfaces on top of Hadoop and I think we're actually very supportive of the evolution, it's a very natural thing, it's something we've anticipated really probably since the start of the company that this would start to evolve and that the way you get out, the raw data in Hadoop is going to get better and faster. It's still very, very early days, good work by Clodera with Impala, the work by Green Plum, the Hawke interface is pretty interesting and there's others as well out there. I think fundamentally though, I think you've got to sort of say, where does that get us to? So let's say these things are, they work just as advertised, which is, you know, it will happen over time. They really are doing a good job of starting to rebuild the traditional data warehouse stack inside Hadoop and so, you know, who should be worried by that? You know, they think the Teradata's, the Oracle's, the folks who are taking the alternate position which says, you know, we don't use Hadoop, we'll use it as a, just an offshoot of our relationship database. It's trending, people say, I want some Hadoop. It's bolted on and you get some cheap data warehouse. I think the difference for us and why actually we're incredibly, we feel incredibly good about this sort of, this evolution is, it only highlights the need for business value from all these Hadoop investments. That's the missing piece. People, you know, you pour all that data in and now business users crowd around it saying, well, I want to do something with that and before it was, I've got to go get a bunch of programmers and I'm going to spend months working on it to try to get at this data that's, you know, you captured it in the data reservoir. That's great. Now, how do I get to do something with it? You had SQL interfaces, now it's a little different. Now you can hire DBAs and spend six to 12 months modeling it and building aggregates and mapping it. And there's some infrastructures, racks of gear involved, right? Yeah, and fundamentally you've got all the IT pain of before, now just rolled forward in a way that, you know, if everything aligns and the technology was really, really even better than anticipated, you'd probably, you know, be able to get up and running in six to 12 months, doing a project around this stuff. And, you know, that's great from the existing BI vendors who want to sort of tell a story about being relevant in this new world. But I think that the biggest story is, okay, these things are great, but there's a much different architecture that we're excited by. It's about drawing out and accelerating and building these in-memory accelerated software-defined datamots that are automatically generated from the data reservoir. Software-led. Yes. Software-led datamarts. It's not defined yet. No, no, it's a good point. Okay, datamarts is not a one-trick pony anymore. Absolutely. But I think, you know, the essence of it is get IT out of the loop of doing all that stuff, you know, so that you don't have to wait a year, you can be doing things the next day. And that's kind of where we think this is. So what's your take? So let's break this down, because you kind of went a little general on that. I want to be specific. Let's, EMC, Green Plum in particular, okay. So they're claiming that this is massive performance increase. Okay, so you, your value proposition is very specific. Yes. It takes months and months of schema definition and then, you know, then you're up and running. And then there might be a change. That schema sets up some queries and you get some reports. And then you say, ah, I want to ask another question or do something different. You got to re-architect the schema. That's not a good scenario. No. Is that what Green Plum is proposing? So Green Plum is, they're taking Green Plum database and Hadoop and HDFS fusing them together in a way that you can now use Green Plum SQL to get at that data in Hadoop. And the same problem that you're referring to. But it does. I mean, all of these do, this goes back 25, 30 years. You know, the problem with that is, you know, so even if you built the most optimized Green Plum or Teradata or any of these systems, if you have 10 terabytes sitting in a table and you users are just writing ad hoc queries against that, it's going to take minutes or hours to respond to these things. But putting it in Green Plum and calling it real-time doesn't magically change that. It, in fact, it gets slower because you're reading raw HDFS now. So you still need to go to all that manual on Tivo. Hold on, hold on, let's back up there. It's slower? Oh, it's slower. They're claiming they're faster. It's slower than the MPP database. Oh, it's got it. I mean, you know, Hive is the easy comparison point. Everybody's like 100 times faster than Hive and that's because their opinion on it, as Sean Connelly said, you know. Jeff, what's your take on what you said? Well, there are performance implications, it sounds like, from your perspective. When you're applying a technology that was not designed to run inside a two-byte environment, there's some performance implications for sure. Yeah, I think, you know, the net of it is, it will be faster than going through Hive, but it's not consistently fast. You ask the wrong question, you get a lot of data, it takes a very, very long time to respond. And so it's walking us back into the traditional data warehouse world, which I think is a- Where's some Hadoop bolted on? Where's some Hadoop bolted on rather than saying, how do we natively use Hadoop and produce a model where we can accelerate and provide consistently fast, you know, subsequent performance. Okay, so let me just frame this up. Okay, so that makes sense. They're kind of suboptimal for the future, but they have an install base that makes sense for them. Yeah, absolutely. Okay, old data warehousing model with some BI efficiencies with the MPP database. Okay, great, that's EMC. Let's go to what you guys are doing that's different. So what now, what are you, now you guys are more of a platform that drives dynamic queries. So just take us through the difference between old way and new way. All right, so picture data in HDFS. You've landed all this data from different silos in your organization in HDFS. Now, you can get at that with MapReduce, with Hive, with Impala, with Hawk, all these different interfaces. Eight or 10 more, right? There's all these different ways of getting at it. They all have their rewards and their benefits and what have you. None of these are, something that a business user is going to be tapped into. So our technology is kind of three layers. The lowest level we're driving, whatever those interfaces are, to optimally automatically aggregate distilled, pull metadata and build out things from that data into a scale out in memory layer that is designed for consistently fast, sub-second performance of interaction. So it's like an intelligent cash that's driving you to pulling out the things that are relevant and evolving them as your needs change and then into a completely web-based exploratory but VR environment. Where the key is that kind of closed loop notion that you have to aggregate and distill for performance. So instead of the old way, like every other version, every other vendor where it's you spend an IT person a year ahead of time did this and if they're wrong or you ask a different question, you're out of luck. In our model, the aggregations, the way the data is distilled, it's automatically changing based on what's interesting and relevant to the users, the business users at the end of the day. So instead of waiting months and having developers, literally we'll go in against any of those interfaces really and be able to be up and running in an afternoon and providing business users with interactive sub-second performance. So what examples can you provide? We've got only two more minutes so I want to drill down because we really like your approach. So what is your success in some of the clients you're talking to? Can you share with some of the people that you're working with since you got a lot of funding, you launched it in the fall? So we're going to be telling later this quarter a big customer story around our product but today most of the beta customers, we're in beta today, we're in GA shortly. Most of our beta customers, they're Fortune 500s and Global 2000, web advertising media, financial services, retail as well as in the federal space. There are companies that have, that aren't necessarily the bleeding edge technology adopters, they're focused on, I want to bring all this data together and fundamentally what's the business value story on that? How do I quickly get answers and not to spend time trying? So what is their business value to you? What do they tell you when they say, hey Ben, we really love this product or hey we didn't like this or like that, what feedback are you getting? So the feedback we're getting is that we're the first product where they're landing data in business, different teams in the organization are landing data in Hadoop and that's the easy part and then they're able to ask questions of that data immediately the next day in ways that business users are understanding and able to drive and that's for them a complete transformation where before it was six, 12 months before they were able to get at that data. The other thing in the traditional BI world, one of the issues was adoption is still, the number frequently mentions about 20% in the penetration and given organization. So the issue is around making the tools actually easy and intuitive enough. So what are you doing to make sure we don't repeat that pattern now in this world? Absolutely, so we started with, from day one, a heavy investment in design. So really, some of the best design and usability people in the world. We have one of our, we have a design advisory board led by a guy named Luke Rublowski who was the chief design architect at Yahoo, did bag check and some other companies. Great team of people around on the design side complimenting the engineering. And then we really started from the ground up with both concepts and also technology, the latest in technology. So where the first BI product in the market that uses HTML5 canvas-based technology, that means we're using the latest kind of, the gaming engines in the web browser that's all natively in there. Let's us do hundreds of thousands of marks on the screen, really, really lightning fast. Just, you know, stuff that goes to tablets and phones and the rest very, very seamlessly. And just trying to focus on how do you take a fresh look at this? Build in from the ground up, you know, collaboration, exploration back to the raw data. All these concepts that were never part of the model of these products that were built around Windows and SQL and sort of the old way of doing things. And we can make it just remarkably better. And one last question. So when you're going to market, you know, obviously what occurred to me with the kind of applying the data warehousing model inside of Purdue, that's great news for the traditional BI vendors because they want to sit on top. This gives them a way in. So when you go to a customer, are you typically trying to, are you replacing traditional BI tools? Are they saying we're all in on Hadoop and now we need a BI tool? Or are you living side by side sometimes with the traditional? So our favorite customers, and this is probably the most common case, is they're both feeding to Hadoop because they know that they want to build this data reservoir they're committed. And they've taken a few attempts with traditional BI tools or the sort of first generation ways of driving Hadoop. And they have those arrows in their back and they're feeling the pain. And then when we show them what's possible, it's a very, very rapid proof of value. So the difference is remarkable when you actually see the way you can drive it seamlessly. It's almost like we sometimes think of it like the iPhone versus the old clunky phones before. If you design from the bottom up and get the experience right, it's just qualitatively a whole different league than the other ways we're doing it. My final question as we're getting ready for the break is we've been talking about things old way, new way. That's kind of always like we look at theCUBE because it's an emerging market. Old way, I'll see the EMC Green Plong going to attack their existing incumbent base, try to grope to the future and green plums flexing a lot of muscles. And depending on how the world spins, maybe that'll be right or wrong. But I want you to end with a real description of the folks out there, this new architecture that's emerging. And that's really important because there's a lot of noise, right? So explain what is the table stakes? What's the minimum? What's this going to look like? Sure. Yeah, in fact, if I was to make one prediction for five years from now, the thing that I think more, almost certainly I'd put pretty much any money on is that the idea of the inflexible, the EDW of today, the data warehouse is replaced by the idea of this sort of fluid agile data reservoir that I pour data into Hadoop. And it isn't even about Hadoop. It's the idea that I don't have to make decisions ahead of time. And then the new stack that evolves on top of that that's designed around that, that agile exploratory nature from the ground up. And we aren't going to be the only player, but today we're taking a leading role in making that tangible for people. And I think that that's the biggest shift. That's the shift in big data. That's the big thing that matters and affects pretty much every business out there. All the other stuff, there's a lot of stuff in the fringes, but this is stuff that really is going to change the industry. Well, Ben, thanks for coming on the queue. We really appreciate it. Great to have you on at the end of the last day. And I really appreciate you making the time to come on. This is not an obvious area for the press, the analysts out there to really understand these nuances of the business value of the future. Having really fast solutions is one thing, but actually standing that out with value, business value, which right now is what everyone wants to talk about is important. And we're going to stay on at SiliconANGLE.com. We're going to keep on following this important trend. Stay on SiliconANGLE and keep on following the story. This is the queue. We'll be back with our next guest after this short break. Thank you.