 Hi everybody, this is Dave Vellante, and I'm with Wikibon.org, and this is theCUBE. We're here at day three live at HP Discover in Las Vegas, and I'm here with my co-host John Furrier. Robert Young-Johns is here. Robert is the Senior Vice President and General Manager of HP Autonomy, gave a great keynote earlier this week, has some big announcements to talk about. Robert, welcome to theCUBE. Thank you. Appreciate you coming on. So I had the opportunity to hear your keynote, and we're talking about new types of data, talking about machine data, and you guys announced this thing called Haven. So why don't we start with Haven? It's a very exciting new platform. Tell us what it's all about. Yeah, I think the thing to do is to start by thinking more about what we will mean by big data. And that's something that's exercised us a lot, because big data's in that phase right now where everybody talks about it, but everybody means something different. And we felt we had to do a little bit of work to try to understand and compartmentalize what the big data problem was, which is why we came up with those three categories. The first category being machine data, which is in a way a little bit of a misnomer, but it's things like sensors, it's telemetry systems, which is also all those systems that just spew out data every time you or I interact with them. It might be click streams, it might be the data that Telcos collected as you move between cell tiles, and so on. How do you analyze that data? And the second category we thought of was the more traditional category, business information. Business information is the sort of thing that comes out of your CRM systems, your auto management systems, your ERP systems, and so on. And then the third category, which is the one that personally interests me most from the autonomy angle, was what we call human information. And this is video, text, audio, stuff that we as human beings just create an incredible- Exhaust, we call it sometimes. Exhaust, yeah, every minute of the day. I mean, just walking from here to my hotel, we take a little outside walk, just looking at the number of people taking videos, taking photographs, or just creating fast-forwarding information. So we said, why don't we start there by that categorization, and then let's look at the tools we have and how they map on to those opportunities. Now it turns out, and you can argue this is great luck or great judgment, but life's a lot of luck, often like that, we start with two of the best tools in the industry for addressing the ends of that spectrum, the machine data piece and the human information piece, and Vertica and Autonomy. And we actually began to think that's actually the best place to start from, because most of the traditional big data tools sort of start in the middle, they start with business data, they start with data warehousing, and then they try to extend our voice to these things. I think by starting at the edges and extending inwards, you have potentially a better platform. So that's really where the genesis of payment came from. Next step was to say, if we've got these tools, how do we expose their functionality in a consistent way to other application developers and in fact our own different application developers. And that's where we began to come up with this idea of very clear definition of platform definition of the services of Vertica, Autonomy, and our security partners and ESP provide together with all the connectivity back into the big data sources. Okay, so it's very specifically, it's a platform for big data application development on which people can drive value. So one of the things we talk about in the keyboard all the time is real time. And with mobility in the iPad, it's crystallized the CIOs, hey, I want information now. And search has been around, it's been more of a data mining, you bring a corpus of data, you apply algorithms to it, but now there's always been that latency question. So I want to ask you the platform question because Autonomy seems to be integrating in a variety of different places within HP. How do you guys look at that platform and getting that real time? Because there are technical differences between having a big corpus, applying algorithms to it, popping out some information in the old BI world, but now real time is about speed. Well real time, in a way that is no more because it's not real real time, it's a sort of reduced latency, is the way I think it is. One second. And we actually start with a concept that talks about machine augmented human intelligence, because I think that sometimes when you go to the movies and you see Minority Report, you think we're in a place where we're not. We think your machines can do all the thinking for us. You know, the way we think about it is if you could all mint the human capabilities, you get better decisions. Now how do you go about doing that? You give people more timely access to information, you allow them to iterate queries more quickly. So with Vertico, for example, and it's incredible speed, you could take up massive data set, you can run a query at it, you can get a response in a second, and you can say, hey, that wasn't quite what I wanted. Maybe I look at it this way, or maybe I look at it this way, and you can iterate that query dozens and dozens of times within a few minutes in a way you could never do with a traditional BI tool. You know, traditional BI tool, to me, sort of reminded me of, you know, one of the traps that Professor of Physics taught me about when I started at college, and said, be very careful the way you design the experiment may determine the outcome you get. Yeah. And BI is not like that, the way you set the question may determine the responses. There's a lot of legacy mindset issues, so I want to ask you about kind of an important one today that to riff on that a little bit is that, you know, everyone loves big data, and it's, you know, whatever it means, it feels good, it's relevant, and obviously instrumenting business with data is really relevant, certainly from a business model standpoint. How do you guys view the operationalizing of data? Because now you have a methodology, the value chains of business is how they're instrumenting their business, whether it's manufacturing, supply chain, hiring, hiring, marketing, you guys are offering a variety of vertical applications. So how do you guys talk to customers about operationalizing analytics? Well, I think it has to be firstly built around the use cases, and that's why I come back to this idea of machine learning, to do with intelligence, because in many of the situations you see, you can't just say, here's the tool, I mean, I always pay for that experiment, could I go to fries, could I buy the package off the shelf, and could I go to store, could I get me a big data solution? The answer is probably not. You need people who can help you understand the problem you're trying to solve, which data sources you're going to have to apply to make those decisions. Then how do you have the connectors to do the extraction that's going to be necessary into all those legacy data sources? Then how do you buy the internet analytics on top of that in a sensible way to drive business value? That's the way we think about it. So if you look at some of the, I mean, there's a great example just there here of the social media command center as a good example of working with clients to understand what they're trying to do with big data, and then operationalizing in a very specific way that's orientated to allowing them to understand how their viewers, their fans, whatever it is, are reacting in real time, using that wonderful word, to what's going on in their viewing experience or their shopping experience, whatever it is. What have some of the customer conversations been like for you guys? I mean, honestly, everyone's like, give it to me, give it to me, I want data, I want data. As they apply the big data, what are some of the conversations? Are they kind of elementary conversations? Are there more advanced talks around the implementation of rolling out analytics? What are you, did you share some insight into that? Well, I mean, the conversations are trying, as nothing changes here in the world with IT, it's trying to understand the underlying business problem that they're trying to solve. So, if it's a problem of making a fan base feel more intimately the sport that you operate, that's the business problem. Then how do you go about doing that? That turns into the analytics. It's very rarely the other way around. You know, when it becomes a solution, so it keeps reaching for a problem, that's usually a recipe for a bad logic. Yeah, exactly, yeah. Can you talk a little bit about the whole developer, Miami Haven's a platform for developers? So, how will HP change the way or enhance the way in which it approaches developers? Well, I mean, the way we see it, certainly for autonomy and vertica, is that we will expose all the core functionalities, you know, through standard REST APIs or web services or whatever your right terminology you want to choose here. So, the developers can say, hey, I want to do a contextual categorization type search and it was not a keyword search. I need that functionality. I can talk through this API or this web service. I can get that functionality. You know, it could be, here's a video analysis I need to do. I know that within autonomy, there are some incredibly sophisticated video analysis tools. I can access them through these interfaces. So, our next task on Haven is to be very clear on the documentation and the specification of those APIs, web services, whatever is, so that we can actually then start to get to the point we have a good developer toolkit. And my vision here is I hope that by the time we get to the next discover, you know, we can give the developer kit which will allow people to go, hello world for big data. Yeah, and it's always my test, you know. Yeah, right, so talk a little bit more about that from the standpoint of just developer outreach and how you market and communicate with developers because you're, it seems to me anyway, the effort with an HP has been somewhat bespoke. I mean, obviously autonomy has its, you know, relationships and Vertica has its and other parts of HP. Are you, do you envision consolidating those in a more deliberate approach? I absolutely think that we have to a lot more work to get to energize the developer community. I think what we've tended to do is work with third parties, i.e. companies at OEMR software, or we bridged their analytics to our back end processing capabilities, but this is more about talking to developers. And I think that's the piece of work that we are, you know, starting to think about how we do that across the whole of HP software and this is a good forcing function for it with Haven because it's a beginning of a, this is a big journey, it's a clear vision and we can talk to developers about it and they're all going to ask us now, so what's the next step, how do we engage with it, we need to be able to respond. And I can begin using Haven today, is that correct the platform? There's already functionality, absolutely, and we already have people writing Haven-based applications. I think taking it to the next level is just making it so the generic developer who came to a developer conference could walk away and do something as opposed to having to get to know us very closely. And how do I consume Haven? Do I just go swipe or credit card in the cloud? Is that sort of where you're headed eventually? That's certainly where we're headed, I mean, and my ambition, certainly for the autonomy components of Haven is that all of the services are exposed as cloud services. It doesn't mean to say people will always use them that way, some cases because of security, regulation, and so on, people want to take them on premise, but in general it's got to be a starting point for us. I think George made that point in his keynote, is that he wants the differentiator of HP software to be, you want to consume our software any way you like in the cloud, out of the cloud. Really the decision between the two is a business decision, not a technology decision. Where does Haven live inside of HP? I know it's a little bit of internal plumbing, but I'm curious, where does it wait? It lives inside HP software. Yeah, okay. It's the HP software creation. His genesis was with the Vertica team. He did a fantastic job. He didn't take a lot of selling to me to say, hey, I want to be part of this. Yeah. And very quickly spun up from that. That is a great example of how, even within a large company like HP, when we get a good idea, we could spin on a dime and go make it happen. Robert, I want to ask you about some of the new things that are going on with big data. So I'll see a good perspective on legacy and the mindset, but there's new things that are emerging out of the new data. So it's new use cases, new tooling, and whatever. Also, you are the first HP executive in the history of the company to actually have a hashtag during your keynote, one of your marketing guys. Your branded hashtag, which went viral actually. So I bring that up only in tongue and cheek because that's a folksonomy and you're seeing hashtags, Facebook's rolling it out. And it's just an example of how people are changing their user experience. It's becoming a navigation opportunity, a channel. It's not a taxonomy, it's more of a folksonomy. So that brings up points to that new things are being invented by people seeing the data and having the data and having access to the data. So with that, what have you seen and what are you seeing for new tooling, new capabilities, just anecdotally, and it doesn't have to be specifically customers things, but just from a vision standpoint, this is an opportunity for developers, for HP, George Kedipo was talking about that aggressively yesterday and it seems to be quite intoxicating on one level. So what's your thoughts on that? It is actually, it's a really interesting question. It's one that's actually perplexed me my entire time in the IT industry because the traditional approach to product development is you go out there, you have a customer focused group, you listen to what customers want, you go back in a dark room and then you say, that's what we've got to develop. And the IT world's never been like that because often people don't know what's capable or what tools are capable of doing until someone demonstrates it to them. So they didn't know they needed that. So I think in the world of big data we're very acutely in that case. So I'm encouraging my team to actually go out there, innovate, demonstrate, create stuff. We don't even know what its commercial value is. Show it to people and then just... Iterate. Iterate and produce new ideas. So a couple of examples of that. I was in our Cambridge office a few weeks ago and I was looking at some of the work they were doing on news aggregation and I go, you know, that's really, really cool. The problem is I don't know that anybody knows that they need that yet. So why don't we just put it together, get a website out there, get an app out there and just let people play with it. So we put up a website yesterday actually called news-spectrum.com which does exactly that. It allows you to track news stories in time in a very visual way and it does all the news aggregation for you. Now I'm hoping that people will see that while that triggers everything. And also, conversely too, yeah, Steve Jobs always said, sometimes the product guys have to imagine the future first and put that out there and that's what he's done because the customers, they see things today in context of their view. And so there's an opportunity there. So how do you guys spark that creativity? Are you encouraging more development, more R&D, more outreach, open source, ecosystems, all the above? Well, I'm encouraging the developers to really take this on and I actually had a very amusing conversation with the head of my development organization in Cambridge a few weeks back. Said, you know, I need you to go out and just create a group of people and all they do is fun stuff. And they're breaking the envelope and moving it forward. And he's a very thoughtful guy. He never reacts very quickly because he always thinks things through before he responded. And it was a sort of five second pause and he came back and looked at me deadpan and said, but Robert, everything we do is fun. You know, it's not quite what I meant, you know. It was, you know, trying to just get this idea that there's some stuff we don't actually know what's going on. And the good news is with the cloud and you can do things fast. And you can iterate, option to abandon is very acute and short. So if you see something that's not working, you can just abandon it. Yeah, I mean, our Erasmus platform, I mean, as an example of an area where we're innovating incredibly quickly and, you know, always one step ahead of what our clients think they might want. In fact, our standover there has been sort of really overwhelmed by people. Every time they see it, there's this moment of, oh, I didn't get that. Now I get it. You could do this with that. In fact, one of our challenges there is just trying to contain it now because every time I talk to a customer, I get 50 new use cases, you know, I've got to say at the end of the day, we've got to go develop some code. So I've got to ask you kind of the provocative question. Kind of, you know, let's take a step back and going to the cloud, so to speak. What is the most amazing thing that you've seen with autonomy and big data? I think the most amazing thing is actually comes over initially as fairly preserved. And that is our ability to build archives on just an incredible scale. It's the largest financial. We have financial institutions right now. And it's all peaky computer science, but the fact that it works, sometimes people can even go, wow, this works. And we can search on it. And we can get relevant, meaning-back, a trillion messages. And they're adding to it at seven to eight billion messages a month. You know, it's interesting. You know, you mentioned the scale question. One of the things I get excited about and talking to folks and us playing with data ourselves on SiliconANGLE, Mookie Bond is that the passion on the developers and the entrepreneurs and product guys has always been great, right? You know, that you mentioned, we're doing it, but the scale at which things can get done now is incredible. And you know, I think that's one thing that kind of blows people's minds is that it used to be, you have to ramp it up. It's like the airplane taking off. Now you can get to 30,000 feet pretty quickly. And so, you know, what have you seen on the scale side that has changed the game? Well, there are a lot of things that include the availability of cloud services that have been provided there. But actually, I also want to make a point, which is that scale is actually hot on people's feet. So a lot of things have gone to scale quickly, but they've required a lot of computer science and computer effort together. You know, creating an archive, this is what I'm so thrilled with, you know, these archives, about a million messages, for sure, no problem. I'll get an AWC, I'll get it going in a couple of days. Now you're starting to draw really big data base, access, machine learning. You're duplicating it, you've got to be in synchronization. You're doing searches, you've got to get the same results out of the two archives, so you can make sure the indices are built correctly. You know, I just look sometimes at this stuff and I know that I'm a geek at heart. It works, it's always like a sense of surprise or something, it's complicated, it's difficult, connections. So did you say in your keynote that it's a digital landfill? Yeah. But you can use it, it's not garbage, it's actually good data. That was a slightly different point. That was the CIOs turning around and saying, you know, it's great, I have lots of big data, the trouble is I have an easy feeling that 90% of it should be digital landfill. I should be able to get rid of it and delete it without worrying about it. Well, compliance, that's the compliance question, right? I mean, that's the boring conversation that needs to be had, is that an important one is that it's compliance is kind of boring and not always sexy, but it's the requirement now, tables, they said, you know, people are afraid to delete things now or store things. So it's like, on one hand, some CIOs tell me, I don't want to store it. And some say, I have to store it, I don't know if I'm going to need it. So it's an interesting dynamic. Well, I used to put autonomy's business as you just said the two words, boring but important. You've joined at a really interesting time because, you know, this notion of a CIO trying to manage information as a liability or as an asset. And with Enron and the federal rules of civil procedure, that became quite an important issue where the general counsel tail was wagging the dog, so to speak. And now big data has put more focus on information as an asset. So how has that changed? Tell me, talk about that a little bit and how it's changed your strategy and what you're doing. I have a good friend who's a CIO and a major financial institution. And he was just doing one of those compute nostalgia conversations that people in my generation talk about. Remember when, you know, we were just saying, you know, just five years ago, if you looked to pass me by my priorities, it would have been penetrating systems better. So, you know, we've listened to that and we said, you know, consistent central policy manager. So that, you know, for example, and it may seem a far-fetched example, but it's actually a real one. If you create a Facebook post inside the enterprise firewall, and it's a Facebook post to another person in that enterprise, that may be from a regulatory point of view a message that you need to store. So how do you know that? You know, you have to have something that traps that, talk to the policy manager, says, hey, by the way, here's a Facebook message. It's created from Jim. Jim's in risk management. We send it to Joe. You know, Joe's on the trading floor. Should I archive it? And we need to have a policy manager that goes back? Sure you should. You know, that's no different from a Bloomberg message. Yeah, so it used to be chat was a problem, or phone was a problem, and now it's this Facebook message. So what about the flip side of that, Robert? What about, you may want to delete that message. How do you determine, you know, if I can defensively delete it? I presume that's part of the policy manager. You can only do it from policy. Yeah. And you know, one of the challenges is that, you know, if you've got to have a policy, you've got to implement it. So many organizations will have a, say, a six-month email deletion policy, but you find out they don't actually do it. But that's a hard problem. I mean, because risk by its very nature is decentralized. It's on my, you know, phone and the cloud, and so that's a really, got to have some magic to go and solve it. You know, as we look at that whole cycle of information management, information governance, we're looking at how the products more or more flow into each other. And you know, what used to be backup actually bridges into archiving, bridges into clouds and storage right now. And you've got to actually work those boundaries in a way that companies haven't typically done before. And we're sort of thinking about that. I mean, we're thinking about device backup becomes mobile data information management within 18 months. So I have a team right now making that transition for us because the old world of device backup ceases to be that interesting to people. So what's on my device anyway? Most of the stuff I care about is in the cloud. Most of the stuff is mobile device management. Essentially, I want to be at access to it from whatever device I have. Right. Right. My last question, Robert, is, so you guys seem to be on this cadence of every, every discover, whether it's June or I guess now it's December in Barcelona, coming out with new innovations. And I presume you're on a similar cadence. What should observers look for from Robert Young-Johns in terms of the division you're running? What kind of milestones, objectives do you want to hit? What should we be watching as signs of progress and momentum in the marketplace? I think in the software business, what you should be looking for is innovation. I'm a big believer that software is not a business. Well, it's true of any business, but I think a software more acutely than any standing still is not as competitive strategy. So you've got to see a accelerating evidence of innovation, new products, whether it's in our information analytics space, whether it's in our digital experience space, which we haven't talked a lot about today, but it's a fascinating part of our business, or whether it's in this information governance space. You know, it's almost like, I challenge my team, you know, we may be in the eyes of HP, autonomy is 1% of the revenue. I want to be 20% of the noise when it comes to the innovation that we're creating for the business. I'm 20% of the revenue some days, and I'm always encouraging George, is that you got a big opportunity to do that. We certainly do. All right, Robert, young John. I'll solve all the toys. Yeah, that's good. I'm an analyst, so I get to say those things. All right, Robert, thanks very much for coming on theCUBE. We're double clicking into HP software today. I'm here with my co-host John Furrier, so we'll be back right after this message. We're live.