 Live from Las Vegas, Nevada, it's the CUBE at HP Discover 2014. Brought to you by HP. Okay, welcome back everyone here live in Las Vegas for HP Discover 2014. This is the CUBE, our flagship program. We go out to the events and extract the signal from the noise. I'm John Furrier, the founder of SiliconANGLE, joining with my co-host Dave Vellante, co-founder, chief analyst of Wikibon.org. Our next guest, Colin Mahoney, general manager of the software group of Big Data. Colin, welcome back to the CUBE again, great to see you. Yeah, thank you. Yeah, it's terrific to be here as always. Last time we saw you is at the Big Data SV event we had. I talked about the MapR relationship you guys had. And just overall the success of HP Vertica, Developer Conference last year was a notable highlight from your success. Give us the update. What's happening for you guys here? You've got the main stage here at HP Discover, showcasing all your stuff. What's the big news? Yeah, so lots going on. The big news, we announced Dragline, our next product release, with some great features as always. Every one of the Vertica releases we strike a balance between new innovation as well as performance gains, et cetera. This one, a couple really cool things, most notably inside it. We're introducing parts of what we call Project Maverick, which is our in-memory capability. We can also do live aggregation. So if you're doing very quick look-ups, if you're a telecommunications company and you want to have your customers check their balance, we can handle that type of workload now. We continue to do a lot more with Hadoop and the Hadoop ecosystem. We now support running Vertica on top of Hadoop. Obviously we made the announcement that you referred to with MapR previously, but we'll be able to do that with all the distributions. So we continue to do a lot there and just generally add to the platform around the whole mission of store, explore, serve information out. One of the things that's been really interesting to watch you guys come into HP since you guys were acquired as a startup is a lot of evolution around Haven, success around the developer focus, around some of your customers using the product, very big names. So give us the update. What's happened between last year and this year? Are you seeing autonomy sprinkled in throughout? He's a lot of the big data solutions. Vertica, are you guys in the same boat getting integrated across the division? What's different this year for you? Yeah, so I think the things that are different this year versus last year. Last year we announced Haven and it was here that we made that public. And what we've been working on is more tightly coupling the engine so that there's better integration so that you can load information from one of the engines into the other vice versa, also so that you can query across some of the engines. So we continue to add a lot of innovation in that area and then of course the end in Haven is the end applications on top. So you can see being showcased the digital marketing hub solution that we have here, the operations analytics solution that we have here, service anywhere that we have here. There's a number of different solutions on top of the platform, if you will, that are being built. There are also things like Idle on Demand now where you can actually leverage this from a hosted perspective, parts of the Idle platform. So there really has been a lot of progress in the year that we've had it and lots of partners standing up behind it too. We've announced over 100 partners who are part of the whole Haven ecosystem right now. You guys, Vertica specifically, is a bridge as I see it, a bridge between the old and the new. You're not traditional enterprise data warehouse and you're not pure play adobe. We just did a survey in the Wikibon community about 300 practitioners. Now we biased the survey a little bit. They had to have knowledge of analytics and big data. And the question was, had your organization shifted any workloads from a traditional data warehouse like TerraData Oracle or mainframe to adobe? 61% said yes, we've already done that. And 34% said no, but we're going to do it within the next six months. So by this year, that's 95% that we're taking resources away from the old, bringing it to the new. So where are you in that transition? You're in the middle of it. So you must be seeing all that action going from right to left to left to right. Yeah, so I think we're in a great spot because what happens is people want to move those workloads. So you might move the information to Hadoop, but you don't want to give up the sequel or the analytic capabilities or the performance that you might have associated with your traditional enterprise data warehouse solution. Or you want to augment it with a bunch of new data, log data, sensor data, other types of clickstream data that is just too expensive to put in the traditional data warehouse. I think that bridge that you're talking about with Vertica is we bring sort of the best of the old world, namely sequel, the ecosystem, the BI ecosystem that still works with the products. But we also address the things that didn't work in the old world, the performance, the scale and a much better price performance ratio. So I think that's where we are a great bridge between the two worlds. And it's an exciting time. I think so many customers are just tired of paying the maintenance fees to the traditional dinosaurs and they're looking for something new. But they're also looking to do workloads that they've never been able to achieve before. Have more people accessing information, ask more questions, get faster answers. And so we are in the middle of that with FlexZone that we introduced in our version 7. You don't even have to define a schema like a normal database. You just load your data in and that helps that bridge as well. So those numbers don't surprise you then? They don't surprise me at all. But it's different than a lot of the marketing that you hear, particularly, well, certainly from the traditional data warehouse vendors, but also the pure plays that are trying to placate their partners that are traditional data warehousing vendors. Now you don't have a problem placating Oracle and Teradata. So what do you think about that? Well, I think, I mean, one thing about the numbers too is that have you shifted any workload? And I think there are plenty of workloads that are shifted. The question now I think people are asking is what exactly is that workload that you can do on Hadoop versus what you could do before? And that's, I think, where our value really comes in because we can actually show that we can deliver those same applications at that price performance. So I think a lot of people are moving data over to Hadoop from traditional platforms. A lot of new data is coming into Hadoop, but now what are you going to do with that information and how can you leverage it? But we're happy. I mean, we're being right in the middle. We're happy to do both. We love Hadoop. We coexist with it. It's the H in Haven. And we're happy to take some of those traditional dinosaur workloads off the dinosaurs and help our customers out. Is that, yeah, you just said it. I was going to ask, is the traditional EDW a dinosaur? I think the traditional EDW, in a bloated sense, the economics just don't work for customers anymore. And the speed doesn't work for customers anymore. Having this batch process where you have all the time in the world to get the data together, it just doesn't, it's not the reality now. So when you look at how expensive those systems were, when you look at the rigidity of those systems, they just don't make a lot of sense. I don't think companies though are going to rip them out of the core. If you have something, you've got some great data models, you run a business with it, you're not just going to rip it out. What you will do is look at the periphery in the edges and say, well, it makes no sense for me to spend all this money over here or over here for that workload. I'm going to move that off or this new workload. I'm going to put it on this new system. And I think over time, what you see is the edge is being eaten around these dinosaurs and ultimately, The edge is where all the value becomes, right? The edge is where the value becomes. You've seen that story before. And then it just moves away. How well do skill sets and the customers translate from that traditional world to the new world? So that's one great question. As much as I love Hadoop, Hadoop is Java and it's been built for programmers. And there's a lot of initiatives now to add SQL on top, including our initiatives. But so many people have, like it or not, SQL has become a very standard language for the BI tools and ecosystem. So translating those skills that people already had before into this new world, you've got to be able to use those same tools. And we do that very well. Hadoop is getting there. It's not there yet. So I think the skills actually translate very well. And a lot of the traditional enterprise data warehouse experts and people that are working on these systems, they also bring a structure, no pun intended, if you will, to how businesses need to operationalize the data, which is a critical component of any type of information and analytics. So when we first met, you gave me sort of a good education on the whole big data and the Duke marketplace. And you know a lot about this space. So I want to ask you what your thoughts are on a lot of the recent moves on the chessboard. You're seeing Intel, Jettison, this distribution, which we're always sort of questioning what they were doing there anyway. Maybe it was negotiating leverage. Who knows that, you know, Intel just putting its foot in the water, spending a few bucks to learn. But you're seeing massive investments in cloud era You know, Hortonworks was mopping up a lot of the partnerships, responding with, hey, cash is not a strategy. Is this all goodness at the end of the day? It was great for media. Is it good for the industry and the customer because of investments or is it a distraction? I think overall I think for customers it is a very good thing. I think you always want to have choice and there's certainly choice in the market, but you also want the backing of that ecosystem to build around it and create some standards. You want it to evolve. You want people to build up a lot of new tools on top and that's happening now. So I think in general it's a very good thing. I think it's a natural progression of any industry. There's some consolidation that happens. There's investment that happens. And I think one thing that all vendors, especially HP realizes that Hadoop and its ability as a distributed file system to have video and audio and text or structured data, it is a great catch-all place to do that. So any vendors looking at the workloads that are running Hadoop and saying more and more of a data center is going to be running that type of workload and that's therefore important. So Meg mentioned Vertica, mentioned Vertica in the last call, maybe the last couple of calls as a growth area, along with a couple other areas that were growing within HP. The tree's not a growth story. We all know that, but there are pockets of growth that are highlighted. You're one of them. What's pumping up that momentum? Is it that you've done personally a much better job of integrating Vertica into the HP system? There's maybe the demand side. It's a big market. What's driving that growth? I think what's driving the growth is there's a perfect storm right now that is big data in my mind, which is massive amounts of information. Unlike, say, business intelligence and analytics 10, 20 years ago where you had to take a small sample size and extrapolate what you think happened, now you can literally look at exactly what happened. We have logs of it. Companies are realizing that in that black sand there's gold. And companies that differentiate themselves with data, that compete with data, they win. They dramatically outperform their peers. Return on investment dollars in big data projects is very high. People talk about 10, 12X, 12 to 1 returns on these types of projects. People want to spend. There's definitely a shortage in the number of people that know how to do this. They're turning to companies like HP that can come with a full solution from hardware to software to services. I think that really is driving a lot of our growth. We asked questions about that. I was, again, impressed with the percentage of customers that said, yeah, the project's really paying off in a big way. It was just an overwhelming percentage. And that wasn't the case a year ago. No, I think a year ago, five years, ten years ago, typically an enterprise data warehouse project would take 18 to 24 months if you were lucky to deploy. And it wouldn't even be clear at the end of that timeframe whether it was a success. And what's different now is that you can very quickly get something spun up and show value. It doesn't have to be boil the ocean, everything in the kitchen sink, but you can show that value and then double down and start investing progressively and leveraging, in many times, the same data for different types of modernizations across the enterprise. I wonder if you could talk about HP's software strategies specific to big data. So I know you don't run the software group, but you're a key part of it. But so as it relates to big data and analytics, Hadoop, how would you summarize the strategy? Yeah, well, and I certainly don't want to steal Robert's thunder this afternoon. But what I would say, and I think you'll hear a lot of this, is big data really permeates everything that we do in HP software. Whether it's on the systems and operations management side, you're getting a lot of log data. Autonomy, of course, with the human information, compliance, e-discovery, digital marketing, Vertica, obviously, with the analytic warehouse that we have, security, right? You can't talk about security without massive volumes of data coming through. And then on the application development side, life cycle management side, it's also all information. Data is changing the way we write applications. Applications are phoning home. Applications are telling us how they're running. They're kicking off health data about what's actually going on in application. All of these are leaving that digital fingerprint, if you will, that is data. And I think that's creating and driving a lot of what we're now hearing about and reading about and seeing and practice around big data. So I would say that there is no software business that is not being driven by data today. It is just permeating everything we do and it really is the software that is the difference between data being data and that black sand and data being gold. Software is what makes it gold. Colin, I got to ask you about the Internet of Things. Obviously that's hot right now, Internet of Everything. Whatever you want to call it, whatever buzzword you use. In the news today, Google is purchasing a satellite maker for $500 million. Everything is connected. The enterprises are looking certainly at this component of sensors or whatever is on the edge of the network. What are you seeing for architecturally with your customers around how they're handling that tsunami of data? If you think about things like satellites, you think about things like that enterprises are bolting on. It could be machinery, it could be manufacturing, retail, on and on. They got to put the data somewhere. What's your take on that? I think the Internet of Things is a great opportunity for all of us. I think having IP addresses on all sorts of devices kicks off that log data. I think the challenge that companies are facing right now is where do they put all that information? How do we ingest, as you said, John, the tsunami of data? More importantly, how do we correlate that with the rest of our business to figure out why does this matter? Should I keep this data around? We have a large cable company here in the United States that basically took the data exhaust, the log data, SNMP data, and they used it for predictive network capabilities in terms of where they should expand their CAPEX. They ended up saving over $300 million in CAPEX, and they increased the service levels for their customers. That was all based on sensor data, log data of how their end users with the set top boxes are turning on their TVs, watching different shows, when they watch them. If you can harness that information, you can really use it to drive substantial either revenue opportunity by cross-selling somewhere or cost savings, as was the case in that example. I think we are just at the tip of the iceberg on the Internet of Things. Everything is going to have an IP address. Everything is going to be able to collect data and send it back. The question is, where are you storing it? How are you securing it? How are you actually helping people understand the meaning of this data? Because oftentimes, you know, Robert Young Johns was joking yesterday, if you try to open up that log data, first of all, you're not going to open it up in Microsoft Word. You're going to get a bunch of text and you're not going to know what it means. So how do you really understand what it means? How do you parse out the key aspects of those logs? And every log might be different. There's no standard for a log. So I think it's just all the issues that folks have been dealing with on the traditional enterprise at Orha's side magnify them out by volume and variety, and that's what the Internet of Things is. Certainly, if you look at all the growth cycles and every emerging great innovation boom, you see manual processes. You see people brute-forcing things. And at some point, there's a tipping point. With big data, we're seeing it with rolling up through analytics, whether it's using SQL, that's been the gateway drug for the enterprise users. Certainly, the tsunami of unstructured data doesn't have transactional elements to its storage. So what's the next thing from your standpoint for you guys to take that next level to bring big data to the modern era, that next level of modern capability? Is it development friendly? All the above, what's your take on that? I think it's all the above. I think from a developer standpoint you've got to make it developer friendly. You've got to make it so developers can build their applications and have a service to the information to do anything they want with it. They might want to do deep data science work. They might want to do simple reporting. They might want to do simple storage and aggregations. They might want to do the live aggregates that I talked about with Maverick. But you've got to make that layer a given. And then I also think for the lines of business, you have to create solutions. You have to package up some of the math and some of the IP so that you don't have to hire a bunch of PhDs in your organization. You can just look at what the data is telling you. And the analogy I always give is GPS systems. There's a lot of data going on in those systems, but ultimately you care about it getting you to the right place and knowing what your estimated time of arrival is. And I think with everything data related, these organizations are trying to figure out. But we're doing our job and making it easy to explore, discover insights, and then once you discover it, operationalizes so the organization can actually monitor and measure against that information. Cal Mahoney, great to have you on the Cube. I'll give you the final word. Share it to folks out there why this HP Discover is so important. What's the big important thread here for this event? Well, from my perspective, it is about mobility. It is about the cloud. It is about big data. It's about security. What I love about HP Discover, especially here in Las Vegas is there are very few companies where you can go from, say, a printer. And by the way, printers are massively important in big data. But there are very few places where you can go from the printer to the PC to the servers, to the storage, to the software, and see how it all fits together. You can walk around here and so many of the demos that you're seeing tie all those things together. And I think what's great this year about Discover is last year we were talking about a lot of these things. We announced Haven. This year you can walk around and you can see how they're stitched together. You can see how that information is collected from servers on a phone home, operationalized in the dashboard so that you can act on it. And more so than ever, I think you can really see that one HP come together with a full stack of services hardware and software. You can see it. Big data is at the heart of the value proposition connecting with the infrastructure, making sense of instrumenting the business. This is the cube extracting that signal, sharing that big data with you. We'll be right back after this short break. Thank you.