 At Big Data SV 2014 is brought to you by headline sponsors WAN Disco. We make Hadoop Invincible and Actian, accelerating Big Data 2.0. Okay, we're back here live in Silicon Valley. I'm John Furrier, the founder of Silicon Angle. This is theCUBE, our flagship program. We go out to the events and extract the signal from the noise. And the event that we're covering is the Big Data Silicon Valley or hashtag Big Data SV. We had Big Data, NYC in New York a few months ago, covered the Stratoconference in New York. And also the Stratoconference here in Santa Clara. We're live at the Hilton in Santa Clara, right across the street from Stratas. It winds down. We're day three of coverage and we're joined here at Pauline NIST. As the general manager, I didn't tell for the enterprise software group strategy. Everything involved, you've been a GM and executive across multiple different functional roles. But great to have you again, CUBE alumni. Thanks for coming back. Thanks for inviting me. It's always fun to talk to you guys. Enterprise software is pretty hot. Enterprise is hot. Intel is always doing great things. Always the bell weather. We love quoting Moore's law until the cows come home here in the CUBE. But Big Data is about computation. It's about systems. And it's so now mainstream in the minds of enterprises. That data is kind of like a term that might not be used anymore because everything will be data driven. We had Joe Heller-Steenon, earlier professor at Berkeley, CEO of Trifacta. And he's like, computer science is upside down right now with the data's at the center of the value proposition. So what's your take on all this? I mean, you sure you got a perspective? Well, I think at Intel the first thing we're really proud of is that Moore's law has really allowed us to enable all of this. Because if you look at the computer, if you look at the storage, if you look at the networking that underpins these huge piles of data. It's really now capable of being done at a price point that lets anybody into the game. And it's really democratized it, which I think is one of the real breakthrough things about it. I mean, just to give Intel some props, first of all, Intel should be recognized as the grandfather of the computer industry. And it's grown up so much. I remember when Intel was doing a lot of investing, they would invest in media back in the 286, 386 days. We're going to invest in the media with these cards and media will provide more computation. So it was kind of, I wouldn't say experimental, but it was obvious trend way back then you could start to see graphics. Now with gaming, with data, there's more and more computation, but now you got cloud, right? Cloud meets big data, data and application developers, it's a perfect storm in this consumerization trend. So how do you guys internally talk about this trend? Because you still got the data center, right? Still data center, you're still not going away. You got the cloud and you got big data. Well, I think the first thing we've done is we have gotten into the big data software business. We introduced our distribution of Hadoop last year, and this year we've introduced our Intel data platform, which is really to create this operating system, if you will, for the data explosion. Because we think it needs to be easier to use. But more importantly, it has to exist in the world that's out there today. I mean, it's very nice to talk about social media. The reality though is that companies get into big data because they want to do more business. And if you want to do more business, you have to figure out how to bring the big data together with your operational business. I think that's why people find the internet of things so appealing because it's a way to really automate everything they know about production lines or devices they ship or what consumers are doing with social. So that I think is the challenge with enterprise, that big data doesn't get to exist in the cloud or anywhere else by itself. It's got to exist with the rest of your world, and that's I think the exciting part, to see how those two come together. So you guys launched your data platform and as I said, operating system for big data analytics, Microsoft has Power BI. The platform wars are continuing, but is it really a war? Is it co-opetition? I mean, how do you guys look at that? I mean, it's not a lot of fight for it, but there is still some confusion. Yeah, I think our biggest commitment to it is open. I mean, we've always been a fan of open source and open systems, and our reason for getting into it is, as you guys know, open source is really wonderful, but when it's left to a loose confederation of tribes to evolve the technology, it takes a while. I mean, I like to tell people, how many years did it take for Red Hat to be profitable? And so if you want to see this stuff get developed faster, we think there need to be some people in the game who are willing to be committers, willing to make the code contributions. And I think we're number two for Linux source contributions. So we just feel that one of the things we can accelerate by doing this is a possessive open source as a way for a lot more people to get involved. Yeah, so let's drill down a little bit into the Intel data platform and tell us a little bit about how that, the role Hadoop plays in that. And just kind of give us an overview of the platform and we can drill into some of the details. So the platform has added a couple of components. It's added an analytics toolkit. And it's added streaming mainly from Apache Storm and Kafka. And it's added spark and shark for distributed computes. Obviously we're a computer company. We want everybody to use lots of computes very efficiently. But then last but not least, we've had an effort in Intel Labs for a couple of years now working on graph databases. And we've pulled that in, so the graph capability will be added to it. But again, with an overall goal of saying, how do we put this together and make it easier for people to turn on? And then because of the commitment to open, we're introducing two versions. We're introducing an enterprise edition that's free. And then a premium edition if you want the service and support and bug fixes and things that enterprises usually want when they go into production. But we think that having the free version is very important to make it easy for people to kind of jump into this. So you mentioned focused and support for open source. So is Intel actively contributing back to Apache Hadoop and what things are you doing in the larger community? Yes, we're active. We've actually probably, since we saw you a year ago, significantly increased our number of committers that we've got working for us and our contributions back to Apache. We are putting a lot of the enhancements we've made for encryption, for performance, all back into the Apache open source. Like everyone, we've got management tools that are ours. But in general, we're putting all of the platform level enhancements that we do back in there for everybody. So let's talk about specifically the real value that Intel adds to Hadoop specifically. So we heard a little bit about bringing security down to the chip level. What are some of the things that Intel can bring to Hadoop that maybe some of the other players in the market can't? Well, a couple of big things. The first one is that we have encryption instructions actually in the chip. So you pretty much get encryption for free. And it still amazes me today whether we're talking enterprise data or big data, why everything isn't encrypted. I mean, every time I see one of these breaches, or I get a note like I was just in Vegas at the Sands and the Sands site is down because they were hacked. Why is there vulnerability? Why don't people just encrypt this? It's in the hardware. It's not hard to do. So we want to make sure that we pull through capabilities like that. The other kinds of things that we do that other people aren't doing is there are a lot of people who have discussions about whether, as John was saying, is big data a special area or does it overlap? And we have a very strong presence in the high performance computing business. And there are a lot of people who think that there's a whole section of big data that's really HPC. And so we've brought in luster into our support for Intel Hadoop. And we're adding support for some of the schedulers in HPC to kind of help cross over this high performance computing and big data world. The range of opportunities has been fantastic and what people in the mainstream kind of aren't getting. They go, Intel's competing with Cloudera and Hortonworks and everyone's got this. But one thing we're seeing out here, I want to get your perspective on, is people are funding their swim lanes. Hortonworks is doing the red hat thing. Cloudera is trying to be an enterprise play and MapR is doing their thing. But Intel is horizontally everywhere, right? So you talk about HPC is a great example. Another one is in-memory, right? So in-memory, I mean, you guys, memory. DRAM's Intel was founded on memory chips. This is real, real wide net here. Could you kind of look, explain the Intel philosophy? And it's not just about the distribution, it's about a bigger picture. It is about a bigger picture and it's enabling, I think, the full value. We've been involved in in-memory for a while because we think that there is a real segment of the market that wants what I call real, real time. And you've heard some, I heard a couple of presentations with it today, but if you've got somebody online and you want to put recommendations in front of them, you don't want them 15 minutes from now. You want them while they're browsing. You want to put connections in front of them at LinkedIn for people that they see. So we've been in memory, you're going to see- Or airplanes that are flying, hey, mountain 15 minutes lag, what? Yeah, yeah, exactly. So you are going to see us take a big step next week with another product introduction that probably puts more memory on this class of system. Than you've ever seen before because we think in memory data, it's here for the masses. That it's no longer what Walmart buys from a couple of proprietary suppliers or what someone, you know, DuPont does to run its chemical lines or, you know, oil drilling people do. You want that real time and you're going to see it everywhere. I mean, you know, if the rumors are right, Microsoft's going to announce a new memory product later this year. So, you know, it's going to have come all the way down the- Well, there's so many new things and the Internet of Things opens up a whole nother Pandora's box. We also had a JaiWire on earlier. They, you know, they were in the Wi-Fi access business and there's also a Geofencing was one standard. Now they have other kinds of patterns that are being recognized. So, you know, if you look at the opportunity, there's many system level things and that doesn't even talk about the data center. I mean, never mind the data center by itself is huge. So machine data is a big deal. Do you guys break that out separately, machine, humans? Well, one of the things we did under our new CEO is we actually now have directly reporting to him a division that is just focused on Internet of Things, our IOT division, to really go after that machine data. Now, it's, you know, it doesn't mean that it all doesn't end up connected to the server somewhere in the server group, you know, or in the cloud organization, because the data has got to go somewhere to get processed. But we just felt that being in, you know, more and more in the low end chip business, everything from the instrumentation to how that data gets sped out. It's something that we think we can add value to. So I think it was kind of interesting to have that reorganization a few months ago and bring a lot more focus to it. And the work we're doing with partners, because we know we're not alone. I mean, we're working with GE. How can you not work with GE, you know? I mean, they own most of the rotating devices in the world to produce the data. Yeah, and they're very bullish. Yeah, Jeffrey Meld and I were on a panel together in Chicago, Mines and Machines. And, you know, there wasn't a bunch of no ops on it. The vice chairman of United, you had people running, you know, oil and gas. Huge, huge things happening. I mean, billion dollars of value, 1% improvement. Well, I mean, it's true for us in our FAB lines. I mean, you know, we had done a lot of this before big data even became a thing because if you have a FAB line go out of spec for five minutes, how many hundreds of thousands of bad parts are you producing? I mean, it's real simple for us to understand the production level value of that as well as, you know, any of these processes. Of course, we have a CUBE alum, Kim Stevenson on all the time. She's proponent of big data. We share some stuff that you guys are doing in there. So, again, you guys are not new to big data. And that's one thing I guess I wanted to want to share the folks out there, but I want to get your perspective on how it's changing the game. You've seen the evolution of the enterprise business and the price software, the big guys Oracle SAP. Now you got open sources, democratize that, flip that upside its head. Now you got the big data fabric kind of coming in. How do you make sense of it? I mean, you know, as an individual, as a geek? One, it's, it all waterfalls down. I mean, that's been true since the beginning of time in this business. You know what, you know, I once walked in with a DRAM upgrade for our PCs and handed it to my husband. And this is very telling about how old I am. He said, do you realize that you just handed me in the palm of your hand more memory than was on the first mainframe we worked on? And that's the story of technology, which is really what big data is doing. It's taking stuff that you used to have to only go to. The five million dollar machines, by the way. Yeah, five million dollar machines and five million dollar software licenses. Okay, and now you're seeing, you know, Hadoop really enable all of this capability. But as I said earlier, at some point it has to come back with transactions. I mean, the only reason that Amazon wants to put all of those suggestions in front of you is so you buy. You know, and I did a lot of work several years ago in my career with Travelocity. And they were really funny because they wanted the lowest cost systems in the world for surfing. Because they said, do you realize Americans will spend two hours looking for a fare that's $25 lower, we're just totally amazed that they will spend that many hours of their time searching to save 25 bucks. But then the minute they click buy, it's got to take a credit card, it's got to be secure. You know, it's got to be a transaction that is absolutely guaranteed, you know, to collect the money. And you have to bring those two worlds together seamlessly. That's the challenge, because you have to somehow translate all of that information that you get out of big data into a transaction that somewhere has value for somebody. Yeah, absolutely. It's really to get the most out of these technologies. It's not enough to just find the insight. You've got to not only provide actionable insight, but actually make that action possible. Exactly, exactly. And that, I think, is what we're still trying to solve. I mean, we have lots of activity on both sides of that equation, and it is going to be in the longer term. And by longer term, I don't mean, you know, next year. I mean, two, three years to see how the platforms evolve. And it does waterfall down. I mean, I think some of the guys that have still got really expensive software out there, you know, ought to be doing some deep breathing. And you know, it's funny because none of the big guys, as far as I understand, have introduced subscription pricing. They have cloud offerings, but you buy your on-prem license and you bring it to the cloud. And you know, that's not going to work for the long term. You're going to see somebody have to fight the bullet and deal with subscription licensing in the cloud. And that's going to see change. What's your view of data in the cloud? We've had treasure data on earlier. It's a full SAS model. They're banking on cloud data. And you know, Nervonix went out of business, and that was a storage company in the cloud. And they gave their customers two weeks to move the data. So you have, you know, which is, I mean, I like treasure data. And I think Nervonix, I like too, but they just poof went out of business. I think that right now, where you're seeing people have faith in data in the cloud, is when they're going to a SAS provider. I mean, you look at somebody like Cerner, who's probably running 40,000 servers now doing medical systems for hospitals. People trust them because that's their business. So, you know, they've opened up this whole mid-market tier of hospitals where there is nothing on-premise. They totally buy it all. But they know Cerner understands HIPAA. They know they understand 24x7. They know they understand replication. You know, is it a low-cost service? No, it's a value service. And I think that's where you're going to see the first inroads for real business data, where people are going to providers. And you're seeing some cloud providers who are really carving out what I call that value niche. You know, people like Virtustream, who are going out there saying, you know, if you want the lowest price, you're going to go to Amazon. But, you know, if you want guaranteed 24x7, you're going to come to us. And it will be a confidence-building thing. I want to get your take. I want to get the Pauline voodoo on the cloud data center convergence from a system standpoint. So share with us your latest thoughts and thinking around the systems design and architecture of the future of the data center. You're obviously seeing the processor thing, ARM and the data center being a potential row. We were at the Open Compute Summit. You saw this homebrew kind of moment, this modern era of hackers kind of going down to the hardware level saying, hey, you know, if you want to build your own, that's an opportunity for Intel, but also it's different. Well, I think, you know, we actually see it, and I do tend to agree with you that cloud is really what's going to drive this architectural evolution. And it's not just the big guys, because I think for a long time now, everybody who masses a lot of these systems wants the flexibility to put the pieces together the way they want to put them together. And the way it works today is you get a motherboard from one of the OEMs where they figured out how much memory goes with the processor. They figured out, you know, how many PCIe slots there are. They figured out, you know, all of those little things, and you can't disaggregate them. And what people would really like to do is buy a rack that's got all those components in, and then if you will, put them together as they want to put them together. And as is always true, frequently you need technology to do that. And there are some things we haven't solved yet. You need things like photovoltaics, you know, to really get this bandwidth and the speed that you need to let people just put... There's a reason the memory is close to the CPU, it's for performance. And until you find some other way to wire that in a way that gives you low latency and high performance, you can't just throw out, you know, the physics principles that work there. But we'll get there. I mean, the technology is, you know, whether you look at the switching or you look at the photonics or you look at all of the evolution, it's going to be, it's going to be like handhelds. You know, you needed miniaturization, you needed the screens, you needed Wi-Fi, you know, there was a reason the Newton wasn't successful. And once you got to the point where everything came together, you could have an iPhone. And we'll get there with, I think, architectures that let people disaggregate those mappings and kind of rebuild them on the fly. And the beauty of that is if you're a cloud provider and you need more computes at the end of the quarter, you know, you can rearrange the way that things are connected and really respond to user demand. And I think that's kind of the nirvana endpoint that everybody's got. But it's going to take some evolution to get us there. Pauline, you want to switch gears a little bit and talk about the competition and how a company like Intel stays nimble. Just before we had on Emil from Neo Technologies, who's got his built on a graph database and also got their graph database. How do you, when you see little innovative companies like that or we're doing some really interesting things and all the players you see over at the showcase floor over there, Shrata, you know, who are coming up with new ways of processing data, analyzing and visualizing it. How does a company, and it doesn't just apply to Intel, but how do a lot of these large companies that have been in the industry for a while, how do you stay nimble enough to compete with those folks and stay on the cutting edge? Well, I think you have to decide what you're investing in for your core business. And obviously, we continue to invest in better tools and modular designs to be able to do SOCs and things with our chip technology. But we're like everyone. We have an Intel capital arm. We do an enormous amount. In fact, I just saw, I think something put by this week that said we were one or two in the Valley for outside investments this year. And, you know, obviously we've got a set of criteria for what we're looking for, whether it's in the big data space or in the server space or just things that we think are fascinating. And so we've got Intel capital out there. Investing in positions, not so much just because like a VC, we want to return on investment. They do get measured that way, but it's more, where do we want to see and fund the technology that we think is going to push the industry in the direction we would like it to go? So I think you've got to do that, you know, whether you have an eventual end to partnering with them or acquiring them or just investing in them so that they stay around. That, you know, the other joke when you walk through the exhibit area across the street is okay, who's going to be here next year? Who's not going to be here next year? You know, who's going to have, I mean, CalZeta, you know, is the most recent example of who, you know, closed the doors. And so I think that that's the challenge, which is who succeeds, who has technology that really adds a lot of value. I guess the question is who do you think will be here next year who won't be here? You know the name names, but categorically, I mean, you think some people are groping for a market position, kind of trying to get some momentum, just might not have the right stuff. Is it the product? Is it outcome-focused? I mean, we're hearing a lot of likes. One guest said, you know, if you're a database, you just fold shop right now. I mean. I think Graf is obviously one of the exciting areas because I just think it's a, in this day and age, a totally different way of, you know, all of us who grew up with rows and columns and, you know, this notion of linking people, you know, by their associations. I mean, I saw some fantastic examples of what you can do with Graf in a couple of talks today, you know, and I think it is gonna revolutionize the way people think about using data. So there are areas like that that I, you know, some people will be acquired, but I think Graf is here to stay. And the question is, you know, that's why we were investing in it in the labs, but I think that you're also gonna see some plays out there for, you know, the smaller companies and whether they get the funding or they get the acquisition is up for grabs. Well, thanks for coming on. I really appreciate it. We didn't get a chance to get Boyd on, but can you speak for him and what his group's doing? Obviously he's out there. Well, he's actually the guy behind the Intel data platform. So he is doing the Hadoop stuff. He's pulled the Graf stuff over with a goal of moving it from the lab to the business side of the equation, working with the HPC guys. Also doing some stuff to, we're investing in a lot of capability and the chip to do monitoring that we wanna pull up and make available to the companies that actually provide those kinds of solutions. So he's got a data center manager product that is doing that, that they're gonna be enhancing. So he's still the king of software for the data center group. Cube alumni as well. Pauline NIST, GM, Enterprise Software Strategy. So what are you working on these days? What's your goals? What's your focus right now? What are you looking at and what's important to you? What are you watching and paying attention to? Well, my group right now is trying, I've moved back, I was straight line to Boyd and dotted line to the data center group and we've just flipped that. And one of the reasons is because I, we're trying to do what we talked about earlier, which is how do you marry big data with the enterprise and how do you make those co-exist in a way that's good for the end user customer. And so that's my, one of my charters right now. And what's your top three priorities in that space? What are the needs of the enterprise that you're zeroing in on? I think it's how we can use technology to bring those two sides together, open of course, because we're very committed to open systems and the continued democratization of that database and helping Intel get growth back, which is all of our jobs right now. Yeah, and everyone's got the new CEO there. So summarize now for the last question, I'll give you the last word is this moment in time right now, big data, Silicon Valley, a lot of action happening, Stratoconference is kind of winding down. What is the key story right now on this moment in time for big data? It's real. I think it's actually moved from being kind of the evolving lunatic thing to something that people recognize that they've got to adopt. They always don't know how, but I don't think there's a business you're in where you don't have to acknowledge it if you still want to be in business a couple years from now. Following this, general manager of Intel, enterprise, software strategy, tying big data to the enterprise, big opportunity, still a lot of growth value, which is still high. I don't see this bubble bursting anytime soon. There might be a little bit of bubble bursting here and there, but for the most part, I totally agree with you. New expectations, people realize big data and data is the center and they have expectations and that's very disruptive. And this is the cue, we are following it here live and Silicon Valley will be right back after this short break.