 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, welcome back everyone. This is SiliconANGLE.com, it's theCUBE, our flagship program. We go out to the events, extract the signal from the noise. I'm John Furrier, the founder of SiliconANGLE. I'm Joe with my co-host, Jeff Kelly here at Wikibon.org, analyst and Big Data doing the presentations, sharing everyone's big report about the market sizing. And we are here with Ankar Gupta, who is the GM of Metascale, here to talk about what's going on here in the Big Data world and analytics, data, infrastructure. It's all coming together as a perfect storm. Welcome to theCUBE. Thank you, thank you John. So here's the update on Metascale. What's happening with the company and what are you guys doing here at the event? And then we want to, let's just get into some of the conversations. Yeah, so John, you guys are really fun to talk to. So I'm glad to be back here. And thank you for having me here. So since last time we spoke about Metascale, we've been doing a lot of fun stuff. And at this event we recently announced our own Metascale branded Hadoop appliances. So what we saw in the market was there is a need for a ready to go Hadoop appliance that can be implemented quickly, comes with fully managed services. The companies do not have to go through the pain point of putting the hardware together, understanding what kind of Hadoop software or Hadoop distribution they need to use and then how to implement it in their data centers. So what's the big trend line that you guys seen that's really lifting up your value proposition here? Is it the data science side, is it the analytics? Is it the Hadoop piece, all of the above? What's the big trend that's floating up the tide? Yeah, it's interesting actually. We see three types of customer base for ourselves. So one is the companies that are really in the early phases of Hadoop or big data. And they're looking to see how to really go about this whole Hadoop journey, how to build use cases, how to build their big data strategy. And that's where our experience from Searsworld where we were born out of comes in really handy. We help companies understand or build their big data strategy. We help them develop the big data center of excellence. So how to use Hadoop, how to build use cases around it and then really build talent, whether you hire talent, you rent talent or you grow it from within. So that's one kind of customer. Then there are others that are in advanced age of using Hadoop. So now they're talking, they already have Hadoop infrastructure and they're talking about how do they put top-notch analytical tool on top of it? How do they put, bring data science on top of Hadoop and really utilize it to manage their data better, to get sense out of the data from different sources that are coming to them. And the third type of customer base that comes to us, which is companies that started on their Hadoop journey because their internal central ITs that said they could do it. But now after making some investment, they don't know what to do about it. And you guys actually published a report on how big reasons why companies fail in their implementation of Hadoop because they don't have talent or they don't build use cases or they don't have an efficient use of the infrastructure. So we see the three big areas. So let me ask you what's changed since in New York we had big data NYC, we talked there but a couple of things that came out is that event was the vertical focus. So your discipline in retail obviously is one obvious one. What other verticals do you see right now in terms of that's really, really going and pushing that analytics to the farthest? Sure. So coming out of retail, it was natural that we have most of the use cases that were tested and used in the retail world. However, because we were born out of a large retailer, it sometimes gets very hard for us to work with other retailers. They see value in what we bring to the table but they may be particular or not very comfortable about sharing data with a company that is from another retailer. So having said that, we're seeing a lot of direction actually, retail companies that reaches out to us and where they don't have a direct competition with Sears. So it's one kind of companies that we talk to. Then there are companies in talking about vertical, in healthcare we're seeing a lot of traction. Seems like now with Obamacare and all kind of new rules that are coming up, management of that data and making sense from the data is becoming more and more important for companies. The third sector I would say is financial sector where again, utilization of Hadoop not just to manage data but also the kind of use cases that we have talked about which is using Hadoop for all of your batch processing, using Hadoop to actually reduce your expensive EDW footprints. So those are the kind of use cases we now see some kind of demand of in financial sector as well. Let's talk about the platform wars that are going on. Obviously everyone wants to have a platform. You guys have platform services as well. What does it take for someone to have a successful platform and what are some of the myths out there that some customers think a platform should look like and or the key success factors. Maybe I should just say that. Let's just go into the platform. What does it take to have a good platform in Hadoop? You mean Hadoop infrastructure? Yeah, Hadoop infrastructure and dealing with Hadoop data. Sure, so I mean this is one big problem we saw in the market which is why we launched our Hadoop appliance. So what happens is there is some kind of tussle or differentiation, I don't know what the right word would be between a BU business unit which wants to use Hadoop and make sense out of the data. And then there is central IT which is obviously busy with their day job and all that is on their plates. So now they are not always going out and building that Hadoop infrastructure and managing it, but BU wants it. So they don't have the IT to build and manage the infrastructure and IT central IT doesn't necessarily have time or resources to do that. So I think that's a key of, now how do you go about it? Do you hire somebody from outside and then have them build infrastructure but we still be in compliance with the central IT rules and regulations and such. And so which is one big reason why we launched our Hadoop appliance. It's actually ready to go Hadoop infrastructure which has three main benefits. One, it comes pre-built with our own reference architecture and our own knowledge of what has been successful in the market in managing Hadoop in a large enterprise. So what kind of worker nodes are more successful? What should be the reference architecture for a name node or data node than whether you should use Intel versus AMD? What kind of processor speed? What kind of hard disk do you use? More dense hard disk do you use? What should the size be? And second is it's a completely 24 by seven managed services. So a company doesn't need to really worry about the, worry about taking care of the infrastructure or actually buy or hire resources from their own IT to manage the infrastructure. It comes fully managed. So it's our job, not only we build and implement the appliance for a company, we manage it 24 by seven remotely. And the third benefit, it's really easy to, it's kind of plug and play, right? We have our own switch so you can, we will help them inject and manage data from their existing EDW in their data warehouse. So all they need to do is have the appliance, buy the appliance of the size that they want, put it in their data center and they're pretty much good to go. So I wanna push on that just a little bit. The appliance model and the Hadoop landscape. So when you think about Hadoop, you think scale out commodity boxes. When you think appliance, you think, I think when people think appliance, you probably think exadata versus maybe one of the first appliances they think. But talk about your view of how the appliance model works or I would say works or doesn't work. I think you're gonna say works because first you're embracing it. But how does it work in the Hadoop paradigm? Sure. And are there any trade-offs when you do the appliance model versus the kind of roll your own scale out? Yeah, so traditionally when you think of appliance, you feel like you're buying yourself just a box. And the box has everything for you and that is it. The way we define our appliance, it is a box, but it is a very flexible box. So one it is, our appliance comes with, so it is our own reference architecture and the hardware that we think fits the best depending on the size of appliance you want. But then from there it is very flexible. So I'll give you an example how we are different from others. We have, our appliance doesn't necessarily come with our distribution. So we give our customer a choice. If they have a choice to go with, whether it's Hortonworks or Cloud Era or Apache or any other distribution that they want to put on top of their appliance, then we could do that and manage it for them. Or we could use one of the distribution that we think will fit into that customer's, for their specific use cases and would work for them. So that's really a lot of flexibility right there. That they don't have to bind themselves to a particular distribution. So if they're looking for more security, they may choose Maypower over Cloud Era Hortonworks or they're looking for, to manage the appliance on their own, to buy Cloud Era Manager and you do it on their own internally. So that's one. Second is this appliance is extendable easily. So if you want to grow from 10 terabyte or 25 to 50 and from multi-hundred terabytes, it's really, you just add either more nodes to the same appliance or you buy multiple appliances that are pretty much plug and play. So I think expansion of that infrastructure is not as difficult as probably what it used to be and you don't really need to, because Hadoop balance the data, what was the case before because Hadoop is pretty robust in managing data across different servers, different appliances essentially. So once you plug and play, Hadoop automatically re-balances the data, automatically does the fault tolerance for you. So I think the benefit of having that ecosystem really helps and then this appliance, as I said, is very flexible coming from us, distribution neutral, management neutral and then obviously capacity neutral. That's an interesting approach. Kind of switching gears just a little bit. You talked a little bit about the stress between IT and the business side when, you know, and there's always a conflict there, right? And there has been for years in any number of areas of IT, not just in around data and big data. But talk a little bit about who is your customer? Is it a business person or is it an IT person? Yeah, so one way or another, we end up working with an IT person, which is great, we love them. So typically the user of our, when we are selling Hadoop or we are selling big data services, the user is primarily someone from business. Because at the end of the day, they're looking for value out of that investment. And generally the value is related to what are the use cases, what am I getting out of it? Which is in the form of either making better sense of data or somehow managing the data in a certain way that's better than what they were doing before. However, a whole lot of time, we actually do work with IT directly. In that case, either IT is trying to reduce their EDW footprint. They don't want to buy those extra data boxes or any other whatever EDW they may be using. Or they're looking for a more efficient way to manage their existing EDW. So in both the cases, I think, even we always end up working with IT, but your user could be either. Well, I mean, I think as we, as the industry is kind of evolving, I think it's a good sign when the business is driving more of the conversation. It's like early in this market, it was focused too much on attack. Unless on, what's the business outcome? And that's the way you kind of, that's the way you stall a market eventually. But so if we start moving that conversation to the business side, I think that's a good thing. So tell us a little bit about what, what's your plans for the, for this year? I mean, it's a, it's a big year in this market. You know, we're expecting big things. We did our market sizing and professional services, a huge part of the market expected to continue that way. What's on your plate this year? And maybe even more broadly, how do you see kind of professional services and the services, managed services market evolving as the market generally evolves? Sure. So we see professional services again. So talking about Metascale first, we grew about over 300% over 2012 and 13. And we expect similar or higher growth this year. We're seeing a lot of demand. Metascale is looking to expand, expand multi-fold. We're hiring a lot of Hadoop resources. If you come across some, please send them away. We'll do. That's, that's one thing. We'll trade you some UI guys for, for some back-end to do guys. I think guys, I'll take them. And, and so that's one thing. I think professional services at some point of time, as Hadoop becomes more and more mainstream, we do see that as more, as more companies, Ambrose Hadoop, they will, they will become self-sufficient in management of that Hadoop. And I think there'll be more tools. So for example, the whole MapRs command line tool is very self-sufficient. I've heard from engineers that they really love using it. So similarly, Cloud and a manager. I think at some point of time, these services, when they are economical enough, then, and then there are other versions that are out there, that it will become at some point of time, nominal for companies to use them within the internal, to manage the infrastructure using their internal IT services. But I think where Hadoop is today, I see significant growth. I mean, you guys probably know numbers better than I do, but we see significant growth in the professional market. We see a lot of companies that in the nascent stage, still thinking about, you know, how to really go about it. The market is becoming so crowded. Every vendor is coming out and saying, my solution is the best, my no-skill database is better than the other and my database is faster than other and whatnot. So we're still talking about basics. We're still talking about, you know, guys, think of a long-term story. Think of what your whole big data strategy is all about. How do you use a big data center of excellence? So you investment that you're making today are for long-term because your data is not going to go smaller from here. It's only going to grow with the digital storm that we all have in our lives. The digital storm, it's happening and I totally agree with you. But I want you to, the final question is we've got a break here. I want you to share the folks out there in your own words. What's happening at this moment in history where I'm big data, big data SVR event here. We have Stratoconference going on as well, behind us, across the street. What's the moment that's the most newsworthy right now in this big data landscape? Yeah, when you mention history, I love to put my philosophical head and really think about it, but you know, it's interesting. I think big data is becoming even bigger. Who thought that Hadoop, which was an open source ecosystem about six, seven years ago was going to be such mainstream and there'll be so much development, conferences and videos around it. So even the buzzword today, I think it's still a lot of companies talking about newer technology. I think the buzzword today will be enterprise data hub using Hadoop to manage all of your data, both structured and structured at a single source of truth. And at Sears and Metascale, we have been talking about it for several years now and done a lot of webinars and write papers on it. So I think that's the key word, big buzzword again using Hadoop as your enterprise data hub and then now adding data science and top-notch analytics on top of it will create some wonders that I guess we'll see some interesting results coming out from, I think in medical field and financial field that we probably have not been able to do before in the past. I think the history is going to show, the history books will show, you know, folks like Amar Awadallah, Michael Olson, the folks at Cloudera, and then the folks at Yahoo who did all that work when no one was watching, right? And then brought that out and made it commercially available. Really created an industry, I think, you know, when all the squabbling gets taken care of and everyone starts growing to the next level, I think we're going to look back on this moment. I agree Hadoop is here to stay, the tooling's being worked on in memory, all the good stuff, the goodness is happening and that's only best for customers to get the insight. So it's all about the data and we'll be right back here inside the queue. We're live in Silicon Valley for Big Data SV event, extension to our Big Data NYC event a few months ago here at Silicon Valley. This is theCUBE, Silicon Angle, and Mookie Won's coverage of Big Data SV. We'll be right back.