 Live from Boston, Massachusetts, extracting the signal from the noise. It's theCUBE, covering HP Big Data Conference 2015, brought to you by HP Software. Now, your hosts, John Furrier and Dave Vellante. Okay, welcome back everyone. We are here live in Boston, Massachusetts for HP Big Data 2015. Big Data Conference is theCUBE, our flagship program. We go out to the events and extract the signal and the noise. I'm John Furrier with my co-host, Dave Vellante. Our next guest is Chris Sellen, who's the VP of BizDev at HP Software, HP Vertica, HP Big Data. What's the name? I guess it's HP Software. Welcome back to theCUBE. Good to see you again. Thanks, John. It's always a pleasure, Dave. Great to see you again. HP Software now. HP is on the wave to split the enterprise group. HP Software, Big Role, Vertica. Really kind of tucked under. Here's the quick two cents of how it's all structured and the whole Big Data team. What's it look like? Sure, well, Big Data is, there will be two companies post-separation, which, and by the way, we're operating a separate companies today as of August 1st, but the separation becomes full and final on November 1st, which is a new fiscal year. But as of November 1st, it will be two Fortune 50 companies, I believe. HP Inc, which is basically the print PC and solutions around print and PC business. And then Yula Packard Enterprise, which is where our business unit will be, which is where Big Data will be. Yula Packard Enterprise is basically what is now HP's Enterprise Group, HP Software, and HP Enterprise Services. And it's almost exactly half and half in terms of revenue. So. Okay, so now, the focus of the conference is now growing up three years. I guess that's, you know, dog years, that's, you know, internet years, that's big now. Take us through what's at this year. I mean, you've been involved with the grassroots since the founding of this conference. It's been DevOps. We're very, as you know, we're very high on this conference. We love the audience of DevOps and engineering and customers. What's the evolution? Take us through where we're at. So I think the evolution, as you know, this has always been an event that's been really focused on customers. And what we've been hearing from customers, and by the way, this is not just through in Big Data, this is true across all of HP and the emerging Yula Packard Enterprise. But we're getting much more business involvement, financial involvement. Big Data's not just an IT topic anymore. It never really was only an IT topic, but more and more and more, it's about what can I do with all of this data? How can I use it to transform my business, stay ahead of my competition? So the new Yula Packard Enterprise Value Proposition is going to be built around four key transformation areas. Becoming a data-driven enterprise is one of them. Certainly, Yula Packard is very much of a heritage in the IT organization, but absolutely moving upstream more and more into helping solve business problems, deploy business applications and solutions. And the area that we just made an announcement today of our startup accelerator program, what we're really seeing, because as you probably know, because I know John, you even used to work for Yula Packard a long time ago, is this is a company that we really have an unmatched partner ecosystem and channel. And what we're really looking to do is expand that into the startup community. So even the smallest, earliest stage developers, we've got the Facebooks, the Etsys, the Twitters, the Zingas all here at the conference. The next generation, Facebooks, we want them to be building on our Haven platform as well. We want to be supporting them, enabling them. So we're really looking to expand the definition of developers and really work with a lot of the earlier stage companies as well. It's interesting you bring up the heritage and kind of how you guys are solving problems with startups and also customers is that one of the themes that's coming out of the show is this whole sequel on the Duke thing, which we all have been kicking around. But the thing is that the customers are all saying ingestion's a huge problem. Yes. And two, sequel is the lingua franca of analysts and stated scientists. And that's accelerating usage and adoption. Absolutely. Can you comment on that? And what does that mean for the ecosystem? What does that mean for the channels? It make it easier to sell, it means adopt. Take us through what that means. Well, it's very consistent with what I was saying earlier. Just saying I want to build a data lake without understanding what kind of value we should expect from the data lake and what we want to be able to get out of the data lake. In other words, we want to be able to, we sometimes talk about a smarter data lake, like a data lake that you can actually ask questions and get answers and such. And you know, so much has been focused on just building a lake these past few years and what, as the business organizations, the business buyers, the financial buyers, the CFO, the CMO, start getting more involved in these conversations, they want to know what they can get out of it. And as you said, sequel is the lingua franca right now. Still, that's what people are trained on, people know, and they know it performs as well. So we've basically put our engine, really not just Vertica, but the entire Haven stack on top of Hadoop to be able to help make the data lake smarter. Well, and that whole sequel to Hadoop was a big tailwind for Vertica when it started to occur. Talk a little bit more about the Haven Accelerator Program. What's the impetus for that? Who's it really targeted to and what do you bring into the table? Yeah, well, we've given away for years now. Thousands of copies of Vertica Community Edition a month. We're signing up thousands of developers to idle on demand. We have these free services. But what has happened is that, and we're doing a lot of hackathons, we had a hackathon at this event. We're doing hackathons around the world regularly. But some of these developers are saying, okay, now I want to go build some real stuff. I want to build a real solution. I want to go after the enterprise. I want to be the next Facebook and I want to build on your platform, but I'm still kind of struggling with, I'm not that big, I don't have that much money. We don't want price to be a barrier. So it's basically a bridge program for those early stage developers who want to really start taking something into production and really building out this next generation of big data solutions, analytic solutions in different vertical markets, horizontal markets, we want to be able to help them and enable them and take the price barrier out of the equation. So we've created a program for them you have to qualify for. It's a year-long program. You do have to qualify. You can apply on the web. It's basically special pricing, assistance, we give support. So again, it's a step above the free tier which we've always had, and we're basically trying to- You're not taking any equity though, right? No, no, it's not an investment program. So the use case is startups who get used to the technology and want to build scalable gen one, right? That's pretty much what you guys are targeting. And is it free, or is there a discount? Yeah, it's basically a special pricing program. If you go to vertica.com slash accelerator, no, it's not free. Some parts of it are free. No, no, actually there is a free tier. There is a free tier. I mean, we've always had free. We've always had community products free. And there's actually, there is a free tier of the startup accelerator program, I correct myself, but there's also paid tier as well. A lot of that is around support. A lot of it is around support costs, because you know, and as you probably also heard us say this morning, hopefully heard us say, we're really doing more and more working with open source. Our new predictive analytics offering is an open source offering. It's a support only. We don't charge a license fee. So we're moving toward a closer to the open source model. We're seeing more and more of a blended model as well. So most of the costs are actually in support. That's the streaming analytics part of Excavator. Is that right? No, it's not the stream. That's the Kafka integration. Okay. So we're talking about different. But this is the distributed R. The distributed R work, because as you probably know, R has become very, very popular in the data science community, particularly with sort of newer data scientists coming out of school, whose parents told them data science is the hottest job of the 21st century because they read it in Harvard Business Review. It'll be data science. These kids are all coming out learning R these days, but R actually natively only runs on a single note. So you can't truly run R on big data sets. Will HP Labs work to basically create an extension of R that we call distributed R, which we've now productized into Haven Predictive Analytics that lets you use the R language in a multi-thread environment and essentially use it on big data. Because one of the downsides of R is you can really only use it on natively, limited data set sizes. And that's, you contribute that back to open source? And we contribute back to open source. Yes, exactly. So I got to ask you, there was a question from one of the journalists out there that was back channeling me. They're watching theCUBE, of course they all do. Kafka. Kafka. Bearish, bearish or bullish on Kafka, pretty much know the answer since I'm just, you guys are pretty much pro Kafka. But the question from the journalist was, what's bearish versus bullish and why? So give me the why Kafka's so relevant. Well, we're bullish because from a capability standpoint, it really solves a lot of problems. It can really serve as a true intermediate layer and help with all of these data ingestion issues you were talking about before in kind of environments where you've got very heterogeneous data sources, different data coming from different places, different formats. It really, it's a very nice architecture. And it's definitely a question you should talk about more with Shilpa. But the biggest thing is, we also hear our customers asking for it and using it. We've got a number of customers that are using Kafka in production. And we've already got this capability out. They had a great, CB had a great description again. He was just straight up, hey, it's the best. It works well, our customers like it. Yeah, exactly. I mean, we have our customer advisory board meeting before this. We talked a lot about it. There's a lot of enthusiasm. Not all of them are using it. There's a lot of stuff out there. It's still alphabet soup or, you know. Don't say alphabet soup because that's now Google. Okay, that's right, that's right. So, see, I've been so busy with the conference I haven't caught up on the news yet. The URL is abc.xyz. They got all the letters cornered. They cornered the market on the alphabet. But if you go to abc.wtf, you go to Bing. Yeah. Anyway, I digress. Yes. Okay, so give us the update on the ecosystem. Okay, channels are big. And again, this is one of those things that we've been watching with you guys in particular and other large multinationals is that ecosystems are great cost of sales leverage for the manufacturer solution provider like HP. With the cloud and big data kind of coming together because this conference kind of, to me always represented that. I know you guys don't call if this may not like my messaging on it, but DevOps meets real world because analytics are a big part of that. So, a lot of the, you know, you see the Facebook's here, you see that kind of stuff. Yep. But what's going to make it easier for the channel? Is it the sequel on Hadoop? Is it some of the abstraction layers? Is it the Haven Architects? It's some of the abstraction layers but it's also understanding the customer's business. We have our big annual partner conference every year in the spring. I think you guys are there as well. And when I was there this year, I heard some partners, the partners who were doing the best were the ones who were engaging in understanding their customer's business. So not just I can make Vertica run real fast but I can use Vertica or I can create Haven solutions to actually solve problems. Customer loyalty, fraud, those different sort of problem spaces and problem sets. And we've created our Haven application framework which we've also announced here to help some of the partners with that as well. But the partners who are really sort of figuring out how to adapt to technology to solve business problems and bridging the gap are the ones who are doing the best. You know, the ecosystem is a very broad thing because they're also data, we have data providers. Now we have data curation companies who are helping companies both understand their data and what's out there. We've got, you know, tamers here, relations here. We've got a lot of our partners in that space. So you see specialism and expert, kind of domain expertise driving a lot of this? Yeah, that's definitely a growing area. We've got the typical technology partners. Then we've got a lot of our embedded partners as well including the start-ups we talked about earlier and all the commercial ISVs that we're embedding. You mean the embedded partners? You mean people who are OEMing? OEMing, MSPing, yes, exactly. And then of course creating solutions that are based on the Haven stack. Well, Colin in his keynote talks about ERP, how it used to be highly customized, how it became packaged apps. And I thought he was going to say we're following the same track with analytics apps but he put a little different twist on it. Highly customized today with some demand for packaged apps but it's not that simple. You've got to have these sort of composable apps. Composable is a good way to put it, yeah. And that's why we call this a framework and we've had lots of discussions whether we should call it a framework or a platform, Haven's a platform but something you can build on. But yes, if we can get you 80% of the way there and take care of most of the plumbing issues and the ingestion issues and obviously the analytics issues but at the same time how I want to see it, what I want to see, what the front end looks like and kind of what the key metrics are. Organizations want to customize that and they may want to partner to do it for them, they may want to do it themselves because my metrics and even my direct competitors metrics may be very different. It's very tied into how businesses and organizations are differentiating cells in this day and age because that's really what big data is about. Analytics becomes the differentiation and we can't package that too much. We've got to leave sort of that last mile, that last 20% able to be customized. So that's really what we've tried to deliver. What's the dev layer, the sort of tools layer that people call it PAS? What's HP's strategy with regards to that? Well, you know, it's a combination. It's a combination of obviously our Haven components, Vertica, Idle, talked about predictive analytics. A lot of the open source stuff, you know, we were absolutely using working with Kafka. As you know, we work with all the major Hadoop distros. We've been doing a lot of co-engineering with Hortonworks around things like work. And then we've also been working with some of our other ecosystem partners. Like a lot of the front end components, some of them we built, but a lot of them we also co-engineered with the folks at Logi Analytics. So we've been working very closely with them as well. And working with our other ecosystem partners also, you know, we've got a very broad set of capabilities and it's sort of a mix and match components model where, you know, you take what you need, understand what analytics you need, look at what front end you want to build. I mean, we really see this as a solution offering that's going to be delivered primarily through services. In some cases our services and then our services channel. So, but it's analytics enabled at the core. And Excavator is an attempt to sort of integrate more of the pieces, is that right? To create more of the solutions. Excavator is the next major version of Vertica. And I know that you have Shilpan's, she'll give you the full deep dive on Excavator. Yeah, yeah, I can't wait for that. A lot of the open source integration, stuff that we're doing, working around Hadoop, working around Kafka, performance improvements and a lot of work to really enable this next generation of business apps. And also, you know, obviously becoming more cloud friendly, you know, certainly we're seeing more and more cloud out there in the marketplace. You know, it was, you still don't see companies dumping petabytes of data in the cloud, but we have some companies dumping pretty significant numbers of terabytes in the cloud. You say so. Well, and the cloud guys are trying to build a data management layer and provide it as services, but they don't have the sophistication that you can offer. And then you talk to your customers and they're saying, no, the problem we have is data has gravity, we can't, we move into the cloud, we try to move it back, we exchange data and our bandwidth costs go through the roof. So that's the whole thing, the cost model shifts around. And we absolutely have, if you really look at a lot of our really biggest, most strategic customers that are running, you know, hundreds of terabytes, petabytes of data, the cloud economics don't work for them. That's what I'm saying, but it's changing. I mean, it is changing. The cloud is definitely becoming a bigger expect for so. Well, especially for the unpredictable stuff. Yes. You know, where you need that elasticity. Yeah, well that actually came up and I won't name the customer because they probably don't want me to, but that was exactly what they were saying was they discovered when they need to burst, the cloud is a good solution. When they've got variable demand, but when it's steady, the cloud is absolutely more expensive and it makes a lot more sense to bring it on premise. And so you're seeing kind of these mixed environments that, you know, we'll use the cloud for the unpredictable versus stuff, but if we kind of know what our needs are going to be them, you know, it's cheaper to have it in house. But UHP don't really care, right? No, no, I mean, ultimately, that's exactly what we're gearing up for is sort of hybrid cloud, hybrid data center. So, yeah, so a lot of what excavators about is about supporting that more effectively as well. So. All right, Chris, outlook for what you guys are doing, next steps, ecosystem, tell me what's happening. Well, you're going to continue to see us do more and more to, you know, integrate Haven and, you know, make Haven more real in terms of our value proposition around Haven has always been 100% of the data, the structured data, the unstructured data, the semi-structured data, analyze it all together, cross-analyze it and then build solutions on top of it. And that's where the ecosystem and the channel comes in. Because, you know, I think the one thing that differentiates HP the most or I would say the two things are the customer centricity, which obviously you're seeing and who's here. And then our partner ecosystem and channel, which is unmatched by any of our competitors. It's certainly a competitive advantage for sure. And being able to enable that ecosystem and those channel partners to be able to deliver these solutions to their customers because we can't understand every business ourselves. We understand a lot of businesses. We certainly understand IT. You've got a lot of businesses that you guys have. Yeah, we've got a lot of specialty partners that are really enabling and activating that channel even more than we have. Well, I'm super impressed. You've got some really good named stars from Silicon Valley all over the world, here in the booths, using the platform. Are we emming it? Yeah, I mean, building it into their software. When you talk about data-driven companies and you look at who the leaders are and truly the next generation data-driven businesses that are disrupting, you know, a lot of very large percentage of them are customers and we're really proud of that. We'll speak to Colin Mahoney about this, but I think composite apps, composite platforms, you see a lot of Lego block designs, as we say. Yes. You guys are on the right track. Thanks so much for selling VP of BizDev that HP software here inside the cube. We'll be right back after this short break.