 Live inside theCUBE, this is SiliconANGLE and Wikibon's flagship program theCUBE. When we go out to the events, extract a signal from the noise. This is our exclusive coverage of HP Discover 2013 and live in Barcelona, Spain for the European version, which is essentially a bigger version, believe it or not, than the U.S. version. A lot of activity here in terms of just interest, et cetera, always fun. I'm John Furrier, the founder of SiliconANGLE. I'm joined by my co-host, Dave Vellante, the co-founder of wikibon.org. And we're excited to have Robert Young-John, the SVP general manager of HP Autonomy, CUBE alumni, second time on. Last time we had a great conversation. Welcome back. Now thank you. Great to be here. You're one of our most expected guests here. Excited to have you on because last time, we could have gone for an hour because you run and involved in the portion of HP that Dave and I love the most in terms of out in the industry, big data. And it's not just big data in terms of like normal, big data. It's a lot of the fun stuff. It's the tech under the hood. It's the applications on the end. It's the sex appeal on the front end. And it's the geekiness underneath the hood. So you have a great job. I enjoy it. Absolutely. And we like talking to you. So welcome back. So my first question, what's new? What's changed since we last met? What amazing things have you worked on and shipped and what's happening? Well, the big thing we've been working on, which I'm really excited about actually, and we'll make the official announcement today in George Kedifa's speech, is we've taken a lot of the functionality we have in idle and we're now making it a service. So we've got idle on demand. So we've taken all these services like video recognition, like voice recognition, like sentiment detection, and we're going to make them available as individual web services. So developers can come to this and they can string these together to produce cool functions. And one we played with actually, which I thought was really cool, is you're probably aware that IBM Watson has done this jeopardy program and they put thousands of developers that is the millions of dollars. Well, we had a couple of developers who in an afternoon hackathon put together an app that's pretty much as good in two hours, using Wikipedia as the info source and then using these chain idle on demand connectors as the way in which the app works together. So you took a couple engineers during a hackathon, okay? And I'm assuming you used some compute power behind it. Oh yeah, yeah, yeah. Some software. There was some software. So that's good leverage. I mean, that's a good ROI, wouldn't you think? I think it's super. I mean, my big vision, which is real different from when we talked last time, is that if we do our job, a year from now, 18 months from now, we have tens of thousands, maybe even hundreds of thousands of developers who every day are thinking about how they leverage idle on demand, the functionality we have there to create great, well-beating applications. So I want to ask you because what's exciting about your job is you get to put your toe in the water of the future and also play with some of the intoxicating viral successes, like we were talking before you came on about Snapchat, a lot of these applications like Uber, these web scale apps are changing the consumer experience and with the consumerization of our world, these are all big data applications. They are all real-time. A lot of the key table stakes of what these apps are building on are real-time communications whether it's a Node.js, axing databases, using the cloud computing. All that is the messaging here. This is HP's kind of wheelhouse. So tell us, what are you seeing in the vision out there for these new requirements? For whether they're startups, whether they're applications, to be a developer in an enterprise or a consumer, what's the key thing that's driving the innovation? Is it the data? Is it just the consumer demand? What are you seeing? I think it's ultimately the data. And it's ultimately finding new ways of extracting value from things that previously you didn't think had value. You know we use this term dark data. Yeah, I'm not sure what it means, but I think it's a good descriptor for a lot of the information that exists it might spin out of your cell phone network, it might spin out of your log files, click streams, and yet nobody really knows what the value in it is. And as we begin to bring the tools to bear, whether it's our tools in autonomy or the tools in Vertica, we can shed light into this dark data and help people to find trends and other things going on that wouldn't otherwise be able to do. So imagine that you can, as we're showing here on one of our demos, you can actually change media coverage of an event in real time based on social media feedback. It's an amazing application. You can change that. This is a bit frightening. The rules of the game based on how people are behaving in the game. And the timing's really good, right? Because there aren't a lot of really good big data apps that you can get off the shelf and they're because they're hard to build. You need domain expertise, you need tech, you need data scientists, and you need tool sets. So the market is really has a huge appetite for this type of capability. My question is, so I guess we get the developer thing, you got to have that, any good software company has to appeal to developers. What do you actually envision selling? Are you going to sell more idle licenses? Is it more vertical licenses? Is it more hardware? Obviously not going to count that in your group, but what is it that the HP software group ultimately envisions selling that's going to drive revenue and really increase the software component of HP as a company? I think it's a hybrid of two things. I think the first is underlying infrastructure. And for example, for idle on demand, we actually don't know what the business model is right now. I mean, I told the team, stop fretting about the business model, go and do it. Fair enough, John and I talk about it all the time. We don't know how we're going to make this. We'll worry about that down the track. Create value, create value first. Absolutely. But then I think we've also started building some real world, heavy weight, big data apps on top. So for example, a couple of weeks ago, we announced an app that we're really proud of called the Digital Marketing Hub. And what this is doing is trying to change the way marketers think about direct campaigns. Because right now what tends to happen is you do a segmentation, you then develop your marketing campaign against that segmentation, and then some weeks later, you look at what the results are, and then it's sort of too late to get back in the cycle. Now we thought with the big data tools we have, and the ability to do a lot of this analysis in almost real time, maybe we could do that segmentation dynamically and therefore get much better return rates on campaigns. So we produced the software, it uses Vertica capability, it uses idle capability. We demoed it to a bunch of marketeers in New York, and I mean literally, the house went mad because this is what they all dreamt of. The ability to dynamically segment their market in real time so the offer presented to any individual user was custom built that user at that time. That's a great example of how you build big data value on top of a set of core tools. Yes, the digital marketing hub announced that, so that's something that a client's going to put in their facility on-premise, and maybe at some point, or maybe today. Well it's a, I mean, yeah, we'll do both. I mean, I'm pretty relaxed about whether people try to absorb these things or services or whether they choose to run them themselves. I think that's a business decision they have to take, it's not a technology decision. I mean, frankly, I would err on the side of a service, because frankly it's easier for us to manage. Yeah, absolutely, absolutely. Here's the choice, you pick it. So Robert, you have a keynote today at 4.30, your presentation, give us a little tease. Just before the cocktail, yeah. I don't know how I got that slot, it was a test. You got to make some phone calls with it, bump up your profile a little bit. Well you're on theCUBE early in the morning, you get the best slot here, Dave and I are awake. No, seriously, so let's talk about what you're going to present, because one of the things that Dave and I were talking about at Amazon, was re-invent when we were covering that event, and then also we had our own big data event in New York City, around Hadoop World, big data NYC was this humanization element. There's a real art and science of big data, and on the art side, it's really more about humanization. Some of these applications are very geeky, which is cool, but what's also happened is you're starting to see the applications or the insights be very, very humanizing, changing human behavior, the societal benefit, just seeing education. These, a lot of big data applications are moving in that, what's your vision on that? What do we need to do to make things more humanized around the data? Well, you know, I'm not sure that that's something that we as the technology infrastructure put as our highest priority, and I'll tell you why, because actually at the end of the day, I think if you're going to really get value out of this, you've got to energize the developer community. It's the developer community who understand the segments in which they operate. It's developers in the healthcare space who are going to say, I know now how I can take these structured data streams that are coming out of all the instruments that exist in our hospital. Here's the mortality rates, and here's all the unstructured information. I can bring that all together now to have a solution that matters and allows people to assess the sort of outcomes people are getting against various treatment. So you're saying essentially, it's early, let's get the platform and tooling in place and let the outcome of that be whatever that is. I'm a big believer, that's how software works. That's how the best software works. You create an infrastructure, you create a set of tools, and if you're smart, you'd get as many developers out there who come into work every day thinking, I'm going to build great code on these guys' platforms. On the tooling front, what are your areas that you're investing in the most right now? Obviously, when you say you tell your developers don't work with a business model, totally agree on that, because it's early, it's a creative process as well. What are you telling your developers, what are you guys working on right now that's a key priority on the software side? It's the thing I mentioned a moment ago. It's idle on demand. It's taking the function of idle, which was a very large monolithic product, breaking it out into services, 500 or so of these services, making those services available as REST APIs so they can be accessed by any developer and chained together to make really smart applications. Now that's actually in some ways the easy bit. The hard bit is making sure the plumbing underneath exists so when someone starts to throw those APIs at really scale data sources, it continues to work. So there's a lot of hard stuff there on service brokers and all that sort of stuff, we have to work through. So I'm going to assume that HPE, I'm going to trust HPE is going to figure that piece out. A few companies that I'm going to trust to do that. HPE one of them, okay, no, seriously. Now, and I know you haven't figured out the business model yet but what are you envisioning of the choices? You can sell a database, you can license those web services, you can sell that big monolithic thing which you're trying to break up. You can develop new apps like the Digital Marketing Hub or sort of all of the above. Is it an all of the above strategy? I think it is all of the above. I mean, I think for a digital marketing hub that's an application that's oriented towards a particular user community. They will pay for that. I think when you start selling idle as a service, we're going to have to find some analog for value and that might be given it's ultimately about data, it might be on the amount of data consumed, it might be on the data object you associate with the services. Use as many services as you like but if you want a terabyte of data behind it, there's a fee, if you want three terabytes, there's another fee and so on. So I mean, that's speculating because we haven't worked through it. Yeah, new licensing models are things that the world is playing around with. I don't expect it to necessarily affect us. But the thing I want to make sure we do is we get early traction and adoption. I don't want people to sit there worrying about the commercial model either from a developer point of view or from our R&D perspective. Yeah, guess get traction first. Let's get, and that's what everybody does in this space. So I don't think it's new news but I think it's important that we do it that way. Well, it's not new news, but if it's something that you're being essentially measured on by Meg in the board, okay, we care about traction. We don't really care so much about HP software as X% of HP revenue yet. We care about traction so that we can get to that next point. Is that the right way to think about it or not? Well, but you know how this works in a large company. You do what it'd be like, but make your P&L. So I have to fund this within my business model. Well, you're a general manager. You have a P&L, right? I mean, you have a, you produce some results. I can do trade-offs and providing I can attract the value. So I mean, in another area, we have an extraordinary high value application that we've built on these tools, which is our so-called digital safe product. Now digital safe is, you know, but face it sounds pretty boring. It's a archive to make sure banks and other institutions can meet the needs of regulators. It's massive scale. That's the big issue. It doesn't operate at, you know, hundreds of thousands of males. It operates at trillions of messages. You know, and building that, which is incredibly high value to those financial institutions, is a way we can monetize our IP and in turn then make the investments in the underlying technology. I mean, you guys are sitting on a gold mine. I mean, obviously everyone in big data can hand wave that. There's no doubt about it. And I like your approach about the developer. So I want to kind of expand on that line of questioning of the ecosystem. So what's your strategy of the ecosystem? Because you need to, you have an ecosystem. I mean, internally I can see you guys handing out the candy to people. And we've been covering HP now for three years with the Q. So I can say to everyone out there, autonomy is definitely infiltrating the different divisions of HP. So that's good, seeding autonomy around. But to the external world, is there an ecosystem? What's the strategy there? And how do people get involved? Well, the ecosystem I think to me starts with, and I've got this sort of focal event that we're going to hold next year. We haven't quite determined the date yet, but it'll be in the first half of next year, a developer conference. And my viewers, people come to the developer conference, they get the developer kit, they get the code they need, and then you have a little contest and see who's come up with the best code that day. And that's the way you start this flywheel moving. I mean, I was really impressed with what I saw at Forced.com. Not so much because of the product they talked about, but the fact they get 100,000 people, many of them are developers, to come to this. You can be real cynical about everything else. Yeah, but they got 100,000 developers there. That's what I dream of doing in this place. Well, if you buy out San Francisco, you can get anyone there. Well, maybe. We don't quite at that ticket level, yeah. Well, we got a little flavor that the Vertica user conference last August in Boston. I mean, that was very impressed with the quality of the customer. Yeah, but that was a user conference. But it was a user conference. It wasn't the developer conference. The developer conference to me is it's geeky, it's got code streams, it's got breakouts where people say, how you do this with this environment if you happen to be using these development tools, how would you access these services, real-world examples that live people to confidence, they can walk out the room and go and produce something that they can then put on the market that's going to change the world. I think you're right. You have to be the steward there. You have to have an event. You have to enable the developers, you have to give them distribution and value. Developers are very fickle, but they're also very loyal. You give them the right tooling, as we know, and distribution. So I pretend I'm a developer. We have the crowd chat application we did a demo last night, and I say, hey, you know what? I want, we're a small team. I want to build on autonomy. What do I do? Well, I think the first thing you'll do is you'll go to www.idolongdemand.com and that will show you all these services we're exposing. So I download some code? Yeah, and it'll show you all the code snippets you need. It'll show you how to bridge these REST APIs into whatever development environment you're using for crowd chat. And you have a marketplace. Can I push my app to HP customers? That will be the next stage of this. We're building that right now. Vertica have already done that. They've done a great job there. We'll probably just follow them. There's no point in recreating it. Interestingly, using some old web OS technology so we're getting some value out of that. Okay, good. You know, that's the cycle you want to try to create here. Yeah, and what's your take on the current developer community? You mentioned Salesforce. Open source is a big part of that community as well. HP's an open stack and a lot of things. What communities are you watching right now on whether it's analytics or, you know, code on the database side or anything in between? Is there any? I'm pretty open-minded about that right now. Maybe this is a mistake, but I'm trying to take a very broad approach because I actually think the functionality we have plays in so many different business segments. You know, if we can do, I mean, trivial examples like sentiment detection or metadata creation out of legal cases, they're going to appeal to a very specific marketplace. So for example, if you're writing software for the legal profession, you're going to want something that can generate metadata from a legal case coming in. And, you know, we have an API that will do exactly that. Yeah, you can just embed that in. Bundle it in. In fact, I had to be told. I had to be stopped using a demo because someone sent me a legal letter. I decided I'd use that as the demo. It was a little bit too close to the law for us to actually put up on stage there. So that's a question a little different. So I struggle sometimes with the market spaces. So you got kind of infrastructure as a service plus, you know, doing Amazon doing its thing with developers. And then you got sort of Salesforce, sort of SaaS minus, right? And then you got the pass market space. So is that pass market space, you know, how real is it? How much of a demand is there for that middle piece? And, or is it going to be a sort of combination of the top and the bottom layers? What's your take on that? That's a really interesting question, because I think the intellectual case for pass is very, very strong, but hasn't yet caught up in terms of the reality of the market place. I actually think that's ultimately where this goes. I mean, you know, it starts with people say, just give me a virtual machine and then I'll do what I like in it. Then what you do is you start abstracting functions that are common across virtual machines and you start to develop pseudo-pass, if you know what I mean, somewhere between I and P and then you report. But you're right, it's more logical to do an independent of platform or independent of applications. And so, you know, the way we're looking at idle on demand is a series of REST APIs that will be pretty much neutral as to which platform you run them on. And then you'll connect that to object store of your choice. It could be Hadoop, we'd probably build our own object store, we actually got some nice ideas to do that with Vertica. We might have a policy manager that we'll have there as a set of services and you'll bring those things together to create an application. I mean, we're thinking about backup for example and backup is just an object store versus a policy manager. So if you create two generic instances there, you can start to bring real cool apps by just matching policy manager to the object store. Archiving is just a special case and so on. Well, and if you can have the data be accessed by those different platforms instead of all the data being locked into a backup platform, in your right, in the past piece, a lot of smart people are betting that that's a big market. I think ultimately that's where it ends up. I mean, I think that, you know, like most of these markets, it goes through evolution and the simple first evolution is, you know, I got these virtual machines, I need to set up, I don't have capacity of my data center, I just want to send them somewhere else. And that's great, fine. But as soon as you find that, you know, if you're writing apps that require media streaming or you're writing apps that require some sort of common functionality, you're going to say, do I really want to write this? Do I want to write my own MPEG encoding mechanism? No, of course I don't. I want to buy it as a service. And that's when you start to get it transitioning from IAAS to PAS. Okay, now, notwithstanding that you haven't had it all, you haven't had it all figured out yet and you got to hit your numbers. But you're from now, where do you want to, where do you want to see, you know, your group, your, what do you want to have accomplished? Well, I build this sort of in a bifurcated way. I mean, I think I need to get the app side of our business where we deliver direct customer value with the applications we've written on top of idle, you know, growing and fulfilling the market opportunity, I know it's there. I mean, compliance, regulation, whatever you think about it, it's a gold mine right now. It's a gold rush. You know, most of these banks, you know, they switch from saying, can I produce a new derivative to can I avoid getting fined? You know, that's a bigger P&L driver than before. So we go, yeah, we could help that. So, you know, I need to get that moving. But at the same time, I think for the long-term effectiveness of certainly what we do and what Vertica does, enabling a developer community is the heart and soul of it. And our measure of our success is going to be the extent to which we can create a developer conference that people are going to want to come to and ultimately, a bit like Salesforce, they're going to want to pay to come to rather than us having to fund it. And you know, people are going to go out there and create, and it's beauty for the system is you're going to have developer who knows, you know, a specific part of the retail market or a specific part of the healthcare market or a specific part of the manufacturing market, we will never know. That domain expertise is key. That domain expertise, they'll do it and they'll create apps based on our platform. And that's, I think, the way that you can actually see this big data trend that we're in right now is being the early phases of what we saw maybe 20 years ago with a relational database. It's a new market. It's a new ecosystem that's worth potentially trillions of dollars being created. And our objective is to be, you know, the infrastructure provider to that. It's exciting. Not to an ambitious or anything. No, I think it's important. I think, you know, the developer focus is really a great message and very relevant given what you guys have right now. It's a great platform. It's a gold mine. And the market's growing like crazy. I want to get the final word in for you to share with the folks that are new to the, that might be new to your area. What's the most exciting thing happening in the world of big data meets applications? Why is this so, why are we all so excited about this? Why is this market so on fire in your mind? I think it's on fire because for the first time people are seeing tools that are going to allow them to expose value in data they're just collecting almost as an ancillary to everything else that they do. So you take click streams, you take sensor data, you take, you know, anything of that sort. Suddenly you can get value out of it in a way that traditional tools could never do. You know, in this field that autonomy works on human information because you can start to extract value from human information, you can again start to build new applications, get new insights you could never do in the past. And those are the two areas where big data is growing most. I mean, I'm just about to go to a different device and I'm going to expose my three cornered analysis of big data into what I call machine data, into business information and into human information. Everybody thinks about business information but actually it's the slowest growing part, it's the best understood part and it's the one where most tools exist. Most of the growth is happening in what we think of as machine data or human information. And the neat thing about our strategy is with Vertica and autonomy, we've got tools that absolutely hit those two spaces. It's beautiful and you know what? Real time means real time, right? So, very exciting. Robert, always a pleasure. I mean, I could sit and talk all day with you. You're a great legend in the community and congratulations, you're running a great part of the business intoxicating really at many levels. The value extraction is so amazing and this is such a good beginning of a massive transformation. This is theCUBE. We'll be right back with our next guest. I'm John Furrier with Dave Vellante. We're here talking big data, getting value out of the data. This is theCUBE. We're streaming the data live here on theCUBE and we'll be right back with our next guest.