 Okay, we are here live, this is SiliconANGLE's theCUBE. This is IBM's information on demand. We're back in Vegas. I'm John Furrier, the founder of SiliconANGLE.com. I'm joined by co-host Dave Vellante of wikibond.org and we're here to bring you two days of action right now at the center of Big Data. We're calling this Big Data Week because we're here at IBM's information on demand and then we're going to be jetting with theCUBE on the team too in New York City for Hadoop World and Strata, which is going to be a lot of the developers and a lot of the emerging companies. But here, the story is all about IBM. IBM's transformation over the past decade has been phenomenal and Dave Vellante, my co-host Dave. What do you think about IBM IOD? John, welcome back to Vegas. I think this is our 15th time in Vegas this year. I'm not sure exactly how many, but it feels like we live here. But yeah, IOD is IBM's Big IM Information Management Conference. A lot of software content, a lot of really system software and expertise coming in around Big Data. And we're going to hear a lot about that this week. Now, John, we were at IBM Edge earlier this year, which was the storage conference, a much, much smaller event. This morning I was in at the Keynotes, listening to Jason Silva and some of the IBM executives kick things off, big crowd. It felt like I was in the Boston Garden for a playoff game. Really packed house, band going, very high energy, lots of data themes. So Dave, we're seeing the transformation of the Big Data world. Hadoop started out in the Emerging Apache Foundation, emerged out from a developer standpoint. And then over the past two years, three years now, our third season cover in the Big Data sector, it's really now mainstream. IBM really puts an exclamation point on that because IOD has always been called, it's been a software information management kind of show, which has been part of the legacy of the computing industry. But now the notion of Big Data is transforming businesses, the discussion of business value. IBM has positioned well, Dave. We know about the storage kind of reconfiguration, the reorganization of storage. You've seen things like Tivoli, we talked about that at IBM Edge. But now we're seeing IBM take this messaging and the positioning of Big Data and integrating it directly into their core products and services. And this is going to be a real boost for the industry and for IBM, in my opinion. So I want to get your take on this and ask you a question. So look at IBM's performance in the stock market, obviously strong. Get your thoughts on that. And relative to Big Data, obviously this is not hype. IBM's not a real big, high-pitched company. When they go for something, it's really as big business. So what's your take on IBM's decision to really amplify and put a stake in the ground with Big Data? Well, the history here, John, as you well know, IBM had the biggest monopoly the computer industry has ever seen, back in the mainframe days. It had 50% of the industry's revenue and about two thirds of the industry's profits. And then it's all very well documented. IBM transitioned really into software and services, services in particular, but also software. IBM made some very important and strategic acquisitions over the years, starting with the likes of Lotus. It bought companies like Cognos and many, many others that we're going to talk about here. And the decision to do that was really to drive value for customers. It was the executive management's decisions at IBM that software really was the main spring of value, how to connect IT infrastructure to business value through business processes, it's really through software. So the company became really astute. You know, one of the best, we've talked about it many, many times here on theCUBE. Oracle, IBM, we certainly put EMC and VMware up there making that transition to a software company. And IBM is a free cash flow machine. Last year, for example, you asked about stock market performance, IBM outperformed Apple in the stock market. That's how well IBM did. Now recently, IBM's been dinged. Stock probably hit close to 220, if not 220. And then it's pulled back to under 200. Largely because people are concerned about growth. Well, there was a lot of euphoria around IBM. People obviously thinking that it would just wouldn't end. And I think this is a really good, healthy correction. You need this every now and then in the market. But IBM is a very stable, very solid company. Leads with services, drags a lot of software. You know, hardware is a piece of the business that's still very important, but in my view, it's secondary to the software value that IBM delivers. You know, Dave, one of the things I'm really interested in, we're going to explore this here on theCUBE, and then we're going to go to New York and hit the emerging segment around Hadoop World and Strata, is you're seeing software being a real differentiator. And if you look at the shows that we've done this year, just to name a few of the notable ones, obviously Hadoop Summit in June, Oracle Open World, EMC World, DM World, SAP Sapphire, you're seeing the big guys really embracing what businesses are looking for. And that is a complete transformation of how they deploy IT and leverage information, in this case data. So big data's top of mind, but cloud mobile and social's been a key trend. And that truly is changing the infrastructure as well as the applications. And Mark Andreessen wrote from a venture capital firm in Palo Alto Road in the Wall Street Journal about software eating the world. And what's interesting to see is that is that the database market, which was once a very boring market, is now on fire. And you're seeing that notion of big data really enabling this. And IBM has roots with DB2 and databases and a lot of history there. And you're seeing again a cyclical trend that we've seen at least three other times in the computer industry. And that is the movement to specialism and specialty databases. Databases enabling a whole new set of applications. And then that moving into more of a general purpose landscape where you're seeing general purpose solutions on top of which starts out as specialty. So Hadoop, all that stuff going on, we're covering HBase, 10Gen, MongoDB, is really about specialism. But now you're seeing that move to software and general purpose-like capabilities. The integration of SQL and NoSQL, structured and unstructured data. And the implications are grand. I mean, privacy, user experience, application speed, et cetera, et cetera. The list goes on and on. Every vertical is impacted by big data. And the root of that is the database, software, applications and user experience. So I'm really looking forward to that this week. And I think that's something we're going to keep our eye on. And I'll say we covered pure data, recent announcement, pure systems. And I'll say with big data here, it's going to be pretty exciting here at IOD. You make a great point. The database market five, seven years ago was kind of really boring. And of course, DB2 and Oracle, they really own the transaction-oriented databases. Yeah, Microsoft SQL Server to a certain extent. But for the really big, hard problems, it's those two. But now all of a sudden this spade of NoSQL activity has occurred. And we're going to talk about that a lot here at IOD and also later this week at Strata. The other thing I want to mention about IOD is, we've seen companies like EMC really transform its marketing. The cloud meets big data trend, and a lot of people have followed that trend. IBM is not to be outdone. IBM probably the most famous tagline in the computer industry, and I tweeted this out earlier today, is think, right? Thomas Watson sort of created that. You remember the famous name plates or stands all over people's desks, think. Think big is the theme of this conference. And there is just a whole sea of information flowing around big data. And the whole Smarter Planet initiative is really a big branding theme. IBM really is number one in that world. You know, I'll take that metaphor and say think, ask a good question, and get the answer. That's what big data is all about. It's asking questions. And we're going to ask questions this week on theCUBE and hopefully get some real big, insightful answers. So I look forward to it. Yeah, good. So John, let's go into our first guest. We've got Tom Deusch, who is a program director for Big Data at IBM. We're going to get right into it. So first of all, Tom, welcome to theCUBE. Great. Great to be here with you. Good to see you. We're just kicking off day one for us. Now you've been here since Saturday. It's been a long week already. How's the voice and the desert air? Hang it in there, hang it in there. So my first question to you is obviously, we're kicking off the show. Obviously a lot of people are buzzed, but comment on your view of IBM's position and big data and what you're seeing. What's the big IOD vibe here for the show? Obviously the messaging around big data. Could you add some color to that? Sure. You know, I view this as a marathon, not a sprint, right? There's a lot of obviously focus on people that are fast out of the blocks, but what we're finding here, and ConocoPhillips was a great example of this in our kickoff this morning, is that these are real enterprise applications that are emerging, and those real enterprise applications have real enterprise attributes around them, right? So the race isn't necessarily to the first person to get to the hurdles, right? You have to clear the hurdles in good form without tipping over other apple carts. So I think what we're seeing here is really a maturation of the space into real productive business outcomes, right? We're moving past playing in what I call exploratory analytics into real production use cases where the rest of the enterprise matters, right? And you're driving real enterprise outcomes, and that's a different space for the big data. Yeah, let me ask you a follow-up on that, because this is interesting, because we've been covering, this is now our third season, we're at Hadoop World when it really kicked off and I've been part of that whole emerging Apache Hadoop community, and history always starts out emerging, right? You have developers, you have some open source in this generation, where you have specialty solutions, and then it kind of gets everyone's attention. And a lot of people have said, oh, cloud, we're Web 2.0 before it, cloud, now big data, or hype cycles, and obviously people talk about the hype cycle, but really this is not so much hype. There's a lot of meat on the bone here, as you mentioned, it's really kind of hit mainstream. Talk about the dynamics in your experiences. Watching it emerge out of the emerging Hadoop community, obviously with cloud, developers are engaged. You kind of have a perfect storm with big data, to share what your observations and experience over the past two years have been. So I think we're early days still, right? The Apache community, not just in the Hadoop space, right? Because big data's more than obviously Hadoop, but the Apache community as a whole has been just, I mean, the track record of doing phenomenal work here has just really been fantastic to participate in and leverage and work with our customers around. But I think what we're finding is that there's still a lot of hype in the market, and one of the things that in the couple hundred engagements that we've done with our customers that I've personally been involved in is there's a lot of bad guidance out there as well, right? So in the rush to keep up with the hype, we see a lot of companies rebranding what they're doing for the last 10 years is big data now, right? And there's one database vendor specifically that if they stamp their unchanged code base that's 10 plus years old, is big data one more time, I'm gonna probably vomit. Folks, that would be Oracle that Tom is talking about. I did not, I'm neutral here. We love Oracle, we're great for radius. So the key is to look at this in what I call the evolving fit for purpose architecture paradigm, which is the answer used to be a database, how could you shrink and beat your data into fitting it, right? What the NoSQL and the not only SQL movement has allowed us is to really look at developing a fit for purpose approach, which is to say we're gonna match the compute problem to the best underlying compute technology. And what comes out of that are areas of being a really, not only a compelling fit, but really breakthrough analytics. And that's really, I think one of the big themes out of this is that we're not talking solely at an infrastructure level, right? That's important. It's really the business outcomes that are driving this and what I've personally seen in the last, I've been in the space now for almost five years now. But what I've seen in the last really 18 months is just transition from exploratory systems to how does this fit with the enterprise, right? And that theme of enterprise is gonna keep coming up here over and over again. So John, I think what I'd say here is that, as this matures and it moves from emerging to more mainstream, the considerations around what it means to be mainstream and what it means to be reliable and what it means to be integrated is going to be a very important next step in the evolution of the technologies. So, Tom, you talked about what we sometimes call big data washing, you know? I remember we had cloud washing, now you've got everybody's gonna get the data. But there seems to be with a lot of the traditional companies, and I'd love to get IBM's angle on this, the notion that, okay, Hadoop is batch, we're real-time. Does IBM subscribe to that? What's your angle on that? Well, so I think there's, again, in that fit-for-purpose notion, right? You use the right technology for the problem at hand. And we, you know, Hadoop has a batch model, or it's how it's thinks. What we're seeing is that starting to kind of morph into bimodal ways of working with it, right? So there's certainly a lot of work happening and more of this is gonna happen in being out on the strata around making Hadoop much more interactive without requiring, you know, a batch only interface. The other thing what we're seeing is, as people start to realize that big data's not just Hadoop, and again, a fit-for-purpose architecture paradigm, is combining in motion real-time analytics, streaming technologies specifically, you know, including our infrastructure streams engine, working with Hadoop, right? So for example, you build your models, you know, with data at rest offline, you then promote those models up for scoring, right? Or you promote models up that actually change or create synthetic data flows, right? In motion, in memory. And then those funnel back into Hadoop systems. And the Hadoop systems then run analytics and low latency off those that then promote models back up. It's actually something that we call adaptive real-time analytics. It's that interplay of truly real-time systems with at rest data systems, whether they're Hadoop or otherwise, that create these kind of very dynamic, very adaptive information flows. And I think what you're going to find is, certainly it's something we're pushing on, but in the industry as a whole, is that, you know, creating a silo, which may be very, very good at something that doesn't help vector actual customer experience or manage risk in real-time or cut down on fraud, is useful, but is going to be essentially a silo that doesn't live up to its potential. So it's how it interacts with everything. So, let me ask you something about Hadoop, because obviously I'm from Palo Alto now, and I've been out there 13 years from the East Coast, and I heard a comment last night having dinner and talking to a bunch of folks, oh, Hadoop, that's only for the people on the West Coast. And the point being that in the real world, New York and other places where, you know, there's some serious mission critical systems, Hadoop is not viewed as like the big data solution. So two things, one, comment on Hadoop's role in the big data ecosystem in your view, and how kind of the normal business IT, CIO, people who are running applications, used to having the cognosis and the DB2s, got a plethora of multi-vendor, multi-app solutions. How does, explain the Hadoop relationship with existing systems? Sure. Well, my customers on the East Coast and in Europe didn't get that memo, so they're making some pretty good progress here. I think there's a couple of dynamics happening, right? So one of the challenges is that there is this balancing act between, you know, our native tendency to want to collect more data, all data forever, and frankly the cost of doing that, right? There are risks in that sort of approach, right? So one of the things that we're looking at is enterprise disciplines around what we call defensible disposal. So hold on to it, preserve it, protect it, actively leverage it for the lifespan in which it's useful. But after that, you want to get rid of it, right? And I think that sort of enterprise management discipline frankly is lacking in the new space, right? There's a lot of emphasis on just collect everything and then magically it somehow adds value forever and that's not as you probably intuitively know the case. The second is that as you move away from certain of the big web properties into more conventional industries, there in some cases is a healthy skepticism about some of these emerging technologies, right? Some CIOs that I work with are less willing to look past the immaturity, less willing to look past a lot of the, what is frankly very siloed nature of these solutions and really demand that their business analysts not have to learn new skills, right? Demand that the people that are doing the information discovery not have to go look for pockets of information here over there and then somewhere else. They're looking for the ability to have that data really be virtualized, right? So the people think in terms of the currency of the business rather than in the currency of the technology and there's I think something to be said for that. Now we're getting a hook from the producers but I want to ask you one final big question. Take us through your vision of how you see the next year unfolding with big data, kind of where we are, what inning of the game are we at and relative to IBM and information on demand the whole ecosystem that's happening around IBM and outside of IBM. Sure, so we're probably in the second inning, right? And it looks like we're going into a doubleheader here, right? So this is just life as we're going to know what's going forward. I think what you're going to see in the next year is a real focus on analytics and real ROI, right? So moving from experimentation, moving from silo deployments into real run the business sort of initiatives where these technologies are mainstreamed and enterprise class in a sense that they really participate in and uphold the rest of the enterprise sort of attributes that existing more mature systems do. Well, in your opinion, that analytics movement live up to the expectations and the hype as I would say that let's say BI and DW didn't live up to that promise. So living up to the hype cycle that we're in right now is going to be difficult, right? I think what's going to happen is and frankly part of what I view my job and IBM's job to do is get customers through that trough of disillusionment, right? And get them earlier on to that plateau of productivity. And a lot of that is knowing not what to do, right? And I can tell you from personal experience I've learned a lot of ways not to do things over the last five years, right? So in addition to the analytics and enterprise character, I think we're seeing a lot smarter project selection and something I've been asked to get. I've got like four talks on it here at the session. So I think it's a lot about knowing what not to do. And a lot of that is good enterprise management disciplines coming back into this space or being exerted in this space for the first time. I do think in some ways the hype is underdone, right? If you look at the ability to transform industries and build, you know, so for example in the financial services work that I've been able to do we're moving from a product centric approach which is a Groupon approach which is I have all these offers, I want to span my customers, right, into a model that says I have a relationship and I will earn money by serving that relationship, right? Sounds simple but it's a fundamental rethink of the architectures and that's an area where the big data technologies have a native play and I think that's actually under-hyped, right? So yeah, that sort of generic buzz, big data makes everything better, right? Everything's got to be big data, that's BS, right? But the ability to rethink about how you do those sorts of transformative initiatives which I think are inevitable, right? As we all have experiences on the consumer space to drive what we expect from the companies we do business with, I think big data is actually under-hyped. It's engagement-based transactions. That's what we're talking about. Yeah, Tom, thanks for coming on theCUBE. I know at times, I wish we'd spend more time and this is something that I've been really thinking about lately, a concept I call data DNA. As data gets more, as more data comes in where was your origin? What's the mix? Is it matches up between two data sets? So it's something that we've been thinking a lot about and looking forward to talking more with you about it inside theCUBE on our next event and keep in touch. Okay, we'll be right back with our next guest after this short break. We'll be right back. This is theCUBE in Las Vegas for IBM's information on demand event. This is Big Data Week on SiliconANGLE TV. We'll be in New York right after this, so stay tuned to us for all the big data action.