 Live from Las Vegas, extracting the signal from the noise. It's theCUBE, covering IBM Insight 2015, brought to you by IBM. Welcome back to day two of IBM Insight, everybody. This is Dave Vellante, Paul Gillan, and George Gilbert, and we're going to put our phasers to stun. Had a few of those in the queue. So welcome back, everybody, day two at IBM Insight. This is IBM's big data conference. IBM doesn't use that term. It's their analytics event. Probably about 20,000 people here. It's very focused on analytics, unlike Interconnect, which is cloud, analytics, infrastructure. This show is really all about analytics. And interestingly, at a lot of these events, well first of all, the first observation I'll make is big tech events are a lot like rock concerts these days. You get a lot of loud music and very cool imagery and the graphics in the IBM keynotes have been absolutely phenomenal. Some of the best that I've ever seen. Jay Porway comes in, he's behind a screen. It's very sort of mysterious and he's not reading. He's got this all memorized. He's very, very solid and he's talking about the vision of the future. The second point I'll make about the IBM keynotes is at many of these technology events, there's a lot of product, product, product just sort of pushed in your face. IBM is choosing really to lay down more of a vision of the future, bringing in use cases, very visionary and thought-provoking use cases. So today we had David Hasse, who was on the Cube yesterday. He's the ultra-cyclist who rode across the country, 3,000 miles in I think eight days and slept literally an hour and a half a day each day. So he probably had 13, 14 hours of sleep throughout that whole journey. Only 200 people have ever finished that race. He came in second this year, cut a full 24 hours off his personal best using IBM Instrumentation, IBM essentially instrumented David Hasse. He was, it was like the internet of Dave and it was really a quite amazing story. So he was up there today, talking about his experiences. We saw Boeing, it was interesting to hear this 100-year-old company head of their analytics saying, look, you can use data, if you have an agenda, you can use data to make your case. One of the old Benjamin Disraeli lies to him, lies in statistics. And she was saying that culturally at Boeing, they're really trying to find the truth. They're trying to be more open-minded, openly recognizing that this 100-year-old state company really needs to open up its mind and data is sort of that enabler. We heard from Shankar Ramamurthy, who also came in the Cube yesterday, big IBM Global Services executive. Then we heard for some startups. Yesterday we had Radpad on the Cube. They're using Cloudant. You don't usually think of IBM supporting startups. There was another startup on stage today called Hothead Games using Cloudant and CouchDB. And then we heard from Doug Baylog. Doug Baylog is the general manager of the Power Division. He's responsible for both the IBM Power and the Open Power Group. So we heard a lot about that, a little discussion about infrastructure and flash. And then we heard from SETI, the SETI Institute, the extra-terrestrial folks that are trying to find remote or life on remote planets. And then of course we heard from Beth Smith, who's coming on later today. She's the general manager of IBM Analytics, talking a lot about Spark, how NASA and SETI are using IBM and using data to go see if there's anybody out there. And then of course we heard from Ed Dumbill, friend of the Cube. Ed Dumbill started the Strata conference, the Strata Big Data Conference, and now is the executive at Silicon Valley Data Science. So really laying out, IBM, a big vision of the future, how you're going to use data, how data is going to transform your companies. Not a lot of product stuff, Paul. Which was somewhat surprising to me. I had expected to see more of that today. What are your thoughts on what you heard this morning? Well, one thing on the startups, I think IBM has a reputation for not being friendly to startups, but they've actually done quite a lot over the last year. They launched a center in New York, a global cloud acceleration program. They're giving $120,000 worth of cloud credits to startup companies. They have a building in Manhattan where they're nurturing these companies. So I think they do support entrepreneurial companies, but it just doesn't get much notice. The thing that struck me this morning was the repetition of the word cognitive, building a cognitive business. Ginny Rometti was, IBM CEO, was interviewed in the Wall Street Journal, a piece that ran this morning, and she made that same point. What IBM is trying to do is get customers to think of themselves as cognitive businesses, as learning businesses. Not just analytics, but actually getting better as an organization from the data that you process. And that's an interesting concept. IBM is the only company I have seen really articulate that idea of businesses being growing and learning. It would be interesting to see if they can define that as a distinctive position in the market. I'm not sure the cognitive is a word that people are going to latch onto all that readily, but the idea of learning and the whole business getting smarter plays very well into their Watson theme and with the whole analytics portfolio as a foundation. So I think we're going to see that word cognitive a lot out of IBM going forward and we'll see if they can define that appropriately to the market. Well, it's interesting, right? I mean, Watson has become synonymous with cognitive. Cognitive is, you're right, it's a term that not everybody uses multiple times a day, but maybe that can work to IBM's advantage from a marketing standpoint. It is different and maybe they can carve out a brand awareness around that. It is a good showcase for Watson. Watson is a learning system and the value of Watson is you can apply it to your analytics data and it will generate questions, it'll generate suggestions, using Watson to analyze the genome and actually find patterns that humans could never find. I think that's a very powerful concept and whether cognitive is the word that describes it, I don't know, but it is the niche that they're carving out for Watson that so far they seem to have to themselves. So George, I wonder if we could get your take on what's happening with IBM. IBM this summer made a big move in the spark. You've likened it to IBM's Linux move, the Steve Mills billion dollar investment in Linux back in the 90s, which actually was early 2000s. 99. 99, okay. And which really took shape in the 2000s and my contention at the time was that was IBM's attempt to moderate into Microsoft's dominance in the marketplace and it actually was quite effective at doing that. I mean, open source was the one force that had contributed to at least the momentum to slow the momentum of Microsoft's Windows 95 dominance. It became an outright success on its own. Yes, and it got a life of its own and it did what Unix was never able to do and you've likened what's going on in the Hadoop ecosystem to the Unix wars and potentially, well, let me ask you, do you see spark as that unifying factor? Maybe you could add some color to that narrative. The short answer is yes, but I want to up level how we got there first, which is, and we talked about this yesterday, IBM's core DNA is solving hard problems, high value problems for a big business and that on the business side generate sort of high margin returns. So what they're doing now is they're taking their analytic DNA, their deep industry expertise and they're fitting these solutions that have outcomes into the workflows of the existing users or end consumers. And what they were very explicit about yesterday is saying Watson is the secret sauce behind it. Now, one technology does not make an entire strategy and so what was interesting is that they're realigning the company around essentially the delivery of business outcomes. And when we talked to any Chosu yesterday, she said something interesting which was that just the way Gerson reoriented IBM around solutions we're sort of going through the same thing that deja vu all over again with all due credit to Yogi Berra where they have taken the deep industry expertise that was in presumably in the Pricewaterhouse Cooper's organization and the rest of the consulting organization and then they're having these go to market units which have these industry solutions. And for the most part, they are not really packaged. What we learned from, I think it was Harriet Freiman was IBM's belief is that we are not going to see in this decade or sounds like the next be there the big packaged applications that we were accustomed to seeing over the last several decades. The key is we're seeing agile customers or in customer organizations need agility and big packaged applications can't deliver that. So the idea is that on top of soft layer they put blue mix and on top of blue mix they put a whole bunch of application services and Watson and so the secret sauce is to compose all those. Now we're at the technical layer, compose all those into industry solutions and the go to market motion is with the deep industry expertise they have in the consulting organization. I think it's an important point that George is making about industry solutions. IBM has done a number of acquisitions recently in the healthcare business, in marketing, in financial services. They're doing a lot of small acquisitions that bulk up their capability to hit these vertical markets and I think that's a smart strategy. Again, it plays to Watson because Watson, that engine can be customized to specific verticals. Well, George, you've noted that there has been a slow motion collapse in infrastructure software pricing. So what does that mean? That means database middleware is under fire. You certainly see this in Oracle, I mean IBM's own business is seeing that as well. I've commented that it's obviously trickled into the hardware as well, probably even more acute in the hardware, kind of underscoring the whole Dell EMC acquisition. So as a result, companies like IBM and Oracle and obviously SAP's already there need to really move up the stack and invest more in applications. Of course interestingly, SAP's moving in the other direction with HANA but that's for other competitive reasons and business value reasons. So when you think about IBM and its vast portfolio of software applications, remember IBM actually back, I think Paul was probably the 70s was restricted from going into the applications business because of its monopoly. And so once the whole wind-tell thing came about, IBM, the cuffs were off and they were able to acquire more software companies, more applications, they of course acquired Lotus and then many, many, many others since then in manufacturing and retail and horizontal applications and vertical applications. So I wonder if George would be comment on that trend generally and then specifically, how does IBM tie all that together? That's a huge challenge. And that's the $64,000 question and it's another way of asking the question that you first asked me that I didn't really answer, which was they see like everyone else the need for an analytic sort of data platform so that we have the common target for all these analytic applications even if they're mostly custom. And when you identify Spark, they really acknowledge for all their efforts in Hadoop, they see it as something that is for the most part not going to run on premise because of its complexity. It'll be in the cloud, but they also see Spark as a more uniform sort of simplified target platform for this type of analytics. And ultimately, the analytics will run partly where the data is, but also partly where it's inexpensive enough to operate it. So there will be some always on-prem but we're going to see a lot more in the cloud and they made it clear that the difficulty of operating like a Hadoop cluster, all the skills that are involved, the fact that it's not a product, that it's like several dozen independent projects, that can't be papered over easily. The, my observation is that for the large companies, IBM, Oracle, we've certainly seen this with EMC. I would say the same is true for Cisco and I think probably also Microsoft. The time between announcement and actual shipment of the vision of the function that's announced tends to be at least nine to 12 months, at least, oftentimes 12 to 18 months. So what's happening is these large companies who are under fire from the disruptors, whether it's cloud, big data, mobile, social, are essentially freezing the market. They're saying to their customers, look, we got your back, we have you covered, and it's actually smart tactic that they use is to pre-announce a vision, so that the customers say, okay, well, we're not really ready anyway. You know, we can't really adopt all this stuff tomorrow. Anyway, we don't have the skill sets, we got other projects going on, we got to schedule this stuff, we're generally in the fat middle, we're not the leading edge, so all right, that's cool, IBM has that. We'll wait for them to sort of roll this out and I think that's very clearly been the case with SoftLayer. We saw SoftLayer and Bluemix announced, I think it was last year at this conference, and at the time it was like, hmm, what's this all about? Now it's finally starting to gain some traction. SoftLayer, the cloud itself has been maturing, still not to the level that you'd like to see with the Amazon, Google, and Microsoft, but it's getting there, and I predict that it will get there, because why, because IBM's investing, but there's still that gap, there's still that lag time. Now, that brings me to the question on Spark. It seems like in the case of Spark, it's different. IBM's leading, so maybe there's still a gap between what they announced and what they're going to deliver, but it feels like that gap is tighter. The ecosystem doesn't like IBM elbowing its way in. It's nervous about that. What's going on in Spark? It seems like IBM's got a laser focus, we're going to talk to Rob Thomas later on today. On Spark, it's got a zillion developers on Spark, and it seems like it's going to take a leadership position in there. In a way, it is freezing the market, but it's actually not freezing the market, it's pivoting the market toward its vision. I wonder if you could again add some color on that. You know how Steve Jobs used to say, if you push on the curve early enough, you can sort of bend the arc of the universe, and that's sort of what IBM's trying to do, but they're not interfering with Spark as it is now. They talk about this church and state distinction, where the one side is Spark as stewarded by Databricks and all the wonderful work they're doing, including the amazing integration of the APIs, so you can essentially work with all this different functionality all at once, which gives you sort of exponentially more power. What IBM is doing with most of their internal development on analytics is they're porting their machine learning, their SQL analytics, their streaming processors, and those they're going to integrate on top of Spark, independent of what the Databricks guys are doing with their similar functionality. So in other words, they're not going to break what's there, they're going to build around it, and in the same time they're saying to customers, if you want a future-proof analytic platform, start with Spark. So it seems like a perfectly viable strategy to me. I mean, it's not a pure open source, Horton works like approach, but it's essentially they're taking an open core and then on top of that building Amazon-like functionality out of that. Is that a right way to think about it? I think actually they're open sourcing their contributions. The Spark event that we keynoted, well, broadcast at Galvanize in- Yes, so let's differentiate between contributions. I mean, I don't think anybody questions the degree of contributions that IBM's making, right? I mean, your point being, IBM has huge contributions to Spark, probably more than anybody, I would imagine. With starting with the machine learning. Right, so- It's all open- So open source and they're giving that back to the community. However, when it comes to things like streaming and other value on top of that open core, IBM, if I heard you correctly, is going to bring in its own proprietary capabilities. Is that right? I believe they're intending to open source that and have it as a basically a richer option- So in most sphere streams, for example, they're going to open source that? That is my understanding. Really? That was part of what we talked about and got into trouble about. Well, we get into trouble. Well, we're going to have the right people on the Q&A today to ask those questions. So you asked Pitchiano at MIT. He asked about Watson open source. He asked at MIT IQ, are you going to open source Watson? And Pitchiano was like, no chance. Wouldn't you, yeah. Okay, so similar question for, like what is IBM's open source strategy? That's really a Rob Thomas question, not really a Pitchiano question. But so we'll find out today, but so you're saying, George, you're going to open source the entire stack, but then how is IBM going to, it's not IBM's playboat. It's not the entire stack, it's the entire analytic engine. And on top of that, then they put their industry-specific pieces. They essentially assemble what is risk management for banking or fraud prevention. As an off-the-shelf capability. Or semi-off-the-shelf. It's not- Semi-off-the-shelf with a sprinkle of little few IBM services in there. Like the main one is Watson. Remember, because what we heard yesterday was, look, there are other companies that are going to have these analytic services, maybe not to the sophistication of their machine learning and their info streams, but the secret sauce behind it that glues it all together was Watson. I think we're going to have to verify today, but- Hence the question about open sourcing Watson. It seems that the strategy is, you open source the engine, but you keep the value at provider. Right. I believe that's what they're doing. So, obviously what they're doing is working, because again, if you quantify this, IBM's last year announced that it had a $17 billion analytics business. Now, I realize that it's kind of a kitchen sink number. They've got their Cognos in there. They've got their information management in there. They've got certain database functionality in there. They've got some consulting dollars in there. So it's bits and pieces that's pulled together. They've got, I think they have some applications in there. They've obviously got their big data stuff, their big insights. They've got the Spark stuff in there, everything, all the services pieces, organized around so-called analytics. That last year was a $17 billion business, which has grown at 20% through year to date, through September. Okay, so that puts it, I mean, you're talking about IBM adding three, three and a half billion in revenue in one year in that analytics business. I mean, so it puts them over 20 billion. It's interesting to juxtapose that to what's happening in the sort of pure play Hadoop world, where, you know, I'm not sure the whole business is that big. I guess it is, but it's not, certainly not, you know, it's not throwing off that kind of revenue. Now, we wouldn't classify as, you know, big data and our big data definition. A lot of the old Cognos stuff, the traditional BI, certainly the Watson stuff would be in there, all the Hadoop big data Spark stuff. So the IBM's number there would be a lot smaller. You can look it up on Wikibon, what that number actually is, and they're still number one, largely because of the services component there. But I guess my point is, if IBM can keep that trajectory, they're looking at a $25 to $30 billion business, certainly well before the end of the decade. I mean, you're talking about in a few years, a $30 billion analytics business. Now, again, you can quibble about the definition of well, is that really big data? Probably most of it is, or some of it is not, as a pure definition. It doesn't really matter. IBM's a leader in that space. They're like an icebreaker within its customer base, and they're just going to suck in a lot of business from their existing customers. Now, that's the other point I want to make. You always hear the big criticism of IBM is, oh yeah, they're selling that to their existing customers. People must understand that 80% of your business, any company is going to be derived, at least most companies, from existing customers. And that's IBM's strategy to service those customers. Yes, they'll probably bring in some new customers, but the bulk of their success is going to come from their existing franchise. And there's nothing wrong with that. Actually, it's a smart play. So I don't criticize IBM for that. I think that the frustration of observers, particularly on Wall Street, is the pace. You know, the pace of the transition is too slow for most people's appetites. But the big question people get is, oh, is Ginny in trouble? I don't think so. I think she got the right strategy. I think she were fumbling and shifting strategies and reorganizing every six months and desperately buying companies and then reorganizing those, that would be a sign of trouble. But IBM's strategy, since she's come on as CEO and reorganized the company, has been pretty steady. And I think that the IBM board will have patience with her because they believe that that's the right approach. And IBM has navigated through these transitions in its history, so. And along the lines where we talk about IBM analytics being sort of a kitchen sink number, for the longest time, people said that about Microsoft Azure. And then it sort of blasted out of nowhere when they put all the pieces together. And I think if you look sort of in time-lapse photography, last year was all blue mix and core blue mix services. Here they talked about differentiating it with Watson. But I think what we'll see next year is some very, very concrete industry specific demos using blue mix and core analytics, but also based on Watson. That's a great point, George. That industry specific, building that industry specific knowledge, which really derives from IBM's consulting business, which derives from the acquisition of PWC years and years and years ago that IBM has turned it to a huge franchise. Building that into applications is a secret weapon that IBM has that many companies can't compete with. That's obviously Oracle strategy. SAP can play that game, Microsoft to a certain extent, but they're more horizontal. And that is a huge, huge advantage for the company. All right, we got to wrap. Next guest coming up very shortly. So keep it right there, but this is day two IBM Insight. This is theCUBE. Go to ibmgo.com for the social digital experience for IBM Insight. There's CrowdChat there, crowdchat.net slash IBM Insight. Join the conversation. Keep right there, there's theCUBE right back.