 Okay, we're back here live in Las Vegas, Nevada. I'm John Furrier with SiliconANGLE.com. This is theCUBE, SiliconANGLE.tv's flagship product. We go out to the events, extract the signal from the noise, and I'm joined by co-host Dave Vellante. We're here with Pauline Nist, Longtime Cube. Welcome back again. You're a rivaling Pat Gelsinger for the most CUBE visits and sessions. We love it. And the top guest honor is Pauline. So you know we love you. Thanks for coming back on. You got to get a trophy. That's the one in Vegas thing that doesn't stay in Vegas. Yeah, broadcast this out to the world. Exactly. So we were actually talking about this last night, Dave and I, and the guys, to doing a SiliconANGLE.tv after dark session. And doing, because there's so much going on after that night, you know, the parties, everything's going on. Dark data. You know. The dark cloud. You would get a whole different perspective on the business after the parties, as opposed to before them. Stream it live, no on demand. If you don't watch, it's gone. So we need to have that kind of vibe for Vegas. So Vegas after dark, SiliconANGLE after dark, special programming. If anyone's got any ideas, just send me an email. We'll get it on right away. Well, welcome back. IBM, information on demand. Also you're at Intel managing all the relationships. IBM, obviously they're no stranger to computing. Right. They're accelerating performance on the hardware side, which is good for business with Intel. But here at IOD, take us through the history. This is our first IOD with Intel. We've been to IBM Edge. Give the folks a backdrop to what is IOD. What is the purpose of the show? And talk about IBM's transition to expanding on their performance with the software. So IOD is obviously information on demand. It's IBM's big software show, which is one of the reasons we're here, because we've been working. IBM's always been an OEM partner of Intel's. They build great system X machines, and now they've got their pure flex systems and their blade infrastructure. But above and beyond that, we view them as a serious software partner. When you look at who owns a compendium of software, a whole software stack that spans all of its way up through virtualization, through operating systems, virtual database, and now big data. I also kind of look at IBM as the adult supervision of the data. Yeah, I mean our last guest is even Slim Oracle. He just like talked about this to another company. Yeah, yeah, but yeah, as you know, I mean, we're thrilled to hear working on it with IBM on their pure flex systems, which is kind of their answer to the exa boxes that another vendor produces. We think anything that lets people figure out what they need, buy it, turn it on, get it into production, get time to value faster is a good thing. So we work with all of these software partners to help them develop by end user solutions as opposed to just hardware, which was kind of the old game. Now it's, how can we help you solve your problem? And I think IBM does that as good or better than anybody else. How about the adult supervision comment? Cause that's a really good comment. Their performance over the past decade on a stock price basis has been phenomenal. That's the direct result of their business acumen as a company. But that also is from the products and the services they offer. And obviously they kind of dumped their PC business in a way and then kind of moved into services. Talk about why IBM's so important relative to that, being an adult supervisor, because that means they're managing a lot of, you know, mission critical applications. They're in all the big accounts. They have history. I think IBM brings two things to the table. You can't argue with the portfolio of analytic solutions. I mean, there's certainly a lot of buzz and a lot of sex about Hadoop. But what do you do with the data? Well, who's got more tools to help you analyze, model, use the data in your enterprise than IBM? And then the second thing that IBM brings to the table is the business side of it, because I think one of the real upheavals that big data is going to cause. And as I, your previous guest just said, we're really only in the second inning is big data is a business solution, not a technology solution. I think anybody who tries to go in and say, geez, you need to have Hadoop, that's the wrong sale. The real question is, how can you be more competitive? How can you be more successful in an environment where that's really a challenge these days? You've got to be out there differentiating yourself to the customer. Who knows how to sell the business solution better than IBM? And if Hadoop and big data are really a business sale to the business, which then turns around and tells IT, this is what I want to do. How do you help me implement it? IBM knows how to do that. So IBM, obviously a big services company. In fact, many people feel they've been, you know, services led for the last several years. But of course, it made some huge investments in software Cognos and many, many others. What's your take? I mean, you've seen the business evolve. You know, well, Gershner's call to make IBM a services company as opposed to a product company. What's your take on the impact of all this big data? Does that increase the need for services? Are we going to see, I mean, services is still probably, you know, two thirds of the IT business. Do you think that will shift? Or do you think it actually escalates, all this complexity escalates the need for services? I think services is going to be there, but it's going to change. It's going to move from the kind of outsourcing, hosting kind of model to cloud, to big data, to solutions as opposed to the mechanics of IT, which is really what services has been up till now. You go to these vendors because you want them to run the stuff, you want them to code it, you want them to maintain it. Sometimes you want to turn it all over to them and just write them a check. And I think we're seeing that with some major moves like GM is going to bring all of its processing back inside now. I mean, you know, this is a cycle. Whatever you're not doing today is what you ought to be doing tomorrow, I guess is the best metric. Paul, and the other thing we've been tracking at SiliconANGLE and Wikibon is, of course, companies are talking about the Software Defined Data Center. We're hearing a lot about software-defined networking. We've talked about software-led infrastructure, really a new take on the infrastructure needed to provide the power, this sort of big data analytics. A lot of flash, a lot of standardized components. What's your take on that whole software-led movement? I think people want the flexibility that the software definition brings. I mean, whether you're running Cod or whether you're running big data, your needs change very often dramatically, very often, very quickly. I mean, you know, the kind of examples people love to quote is it used to be you waited for monthly retail statistics. Now, the Saturday after Black Friday, you want to know how the sales were. Saturday morning, you know, and that Saturday afternoon. In the paper the next morning or online or whatever. And the stores want to understand do I need to do more markdowns? Do I need to do fewer markdowns? What's moving? I mean, right now I was reading the Wall Street Journal this week that the grand panic is ensuing in retail because the latest economic indicators say that consumers are opening up their wallets and all the retailers believe they're already under stock for Christmas. That there's going to be more demand. So think about that rippling through the system and you know, that's the kind of business need that you see out there today. But that's real business. That's not playing around in the lab with something. That's bet your business on it. Have to have the reliability. Have to have the scalability. Have to have the infrastructure to respond. And that's where the software defined configurations can let you augment those configurations on the fly. So you can add capacity, you can add networking, you can add storage without having to reconfigure. Good. What about the information management old school versus the new school? I mean, I say old school, I mean, like just go back five years, even though IBM's been in the database business, they've been doing a lot of that kind of unstructured structure solutions with their databases. But now decision support systems has been an IT buzzword. What's going on in that market of decision support, information management with big data and some of the emerging technologies? How do you view that market? Well, I think it's, there's a little bit of a temptation right now to say, throw the baby out with a bathwater. You know, relational databases are old school. You don't need them because there's all this movement to unstructured and no SQL. And I mean, you only have to look around here at a lot of the partners that have got boosts up. And I think decision support used to be something that the big companies, you know, the Walmart's, the Target's, those guys used to do. I think now one of the vendors has used the term democratization of data. I think what these new tools are doing and what Intel is very happy to aid in a bet is by industry standard hardware, you know, put a pile of it together there in parallel, run these new tools to help you go through the data. But where suppliers like IBM come back in is that the tools that you need to do something with that data in the open source world are just starting to emerge. I mean, when you look at decision support modeling, those kinds of tools, it's going to take, quite frankly, the Apache community, I think a little while to get something that's ready for prime time. And these guys will give it to you today. You know, hook up your Hadoop, hook up your Netiza, hook up your pure data, connect it all together, get the answers to the questions you want. And then, more importantly, once you get the answers to the questions, do the markdowns, do the transactions, turn it into dollars and cents. I think that's a services angle our writers should be on and post it on servicesangle.com because the system integrators are going to love that new concept because it's going to be more work for them. But I want to ask you a little bit about Intel perspective on things. Obviously, Moore's Law, we all love talking about Moore's Law, but big data really highlights a couple of things that are going on. One is, there's a tsunami of data going on in the marketplace, past 18 months more data has been generated than the history of computing that's been documented. They say the next year is going to be double that. So all this is going on. A lot of data is coming in, old data, new data, data mashups. And then vertical markets, you're seeing green field opportunities for new things, new applications that have never been done before. But then all of this is dependent upon essentially compute power, right? So more data means more crunching of numbers. So what's the Intel perspective? Because that's a real, real area. You see startups trying to build new silicon, new solutions, new software. This computing demand, we thought we were good, zilling into the cores on the server. We need more cores, right? Scotty, give me more power, as they would say in Star Trek. But so talk about the Intel perspective around the computing, because the requirements to process more are now here and then they're going to continue to be here. Well, I think that Intel is happy to see this turn into something that gobbles up silicon. After all, that's what we do for a business. So we like software to find networking. We like, we've got Xeons now and virtually all the storage controllers that are out there. One of the big moves obviously is the whole SSD move where you're really getting storage into the silicon arena now, not just rotating devices, which I think is going to be a game changer. I think a couple of years from now, every piece of data in the world is going to be on something that's, the Semiconductor. Woo-hoo! Well, yeah, everybody wants faster access. You still need the pile of backing store and hierarchical store. That's going to be kind of the next frontier of who you buy that storage stack from. But also you're seeing the move to different kind of cores. I mean, we're looking at, we're building our many integrated core, five architecture because we think that there's a certain element of actually big data that if you're really, really analytics bound wants to have cores that can do processing on column or store, processing on different structures of data really, really fast. There's this crossover between the high performance computing world and the big data world that's also happening in some sectors. So we're looking to innovate wherever we can because we think that if you can get to silicon being part of the solution, you can drive cost on, you can drive performance up, which then opens the market to even more new applications. So a key, you mentioned applications, a key to that vision is really the application, ISV is really exploiting that new hardware architecture. Are they, in your view, moving fast enough in that direction? Or is, certainly there's a segment of people. You know, we're going to have AeroSpike on this later this week we had them on. I mean, they're doing it. We know that, you know, doing atomic rights with Fusion IO, all that crazy stuff. Do you see the traditional ISVs doing that and will they, will they have to in order to really drive this movement forward? Oh, we absolutely do. I mean, we have an RZ on course today instead of advanced vector instructions that were done for scientific processing, for vectors. And every database player out there is using them for column restores because what's a column restorer but a vector? You know, and they found that the speed up that they can get from turning those things on is just tremendous. So, you know, people move at different spaces. You know, sometimes it's a little bit of hopscotch-ing in terms of new apps versus the traditional guys. But I don't underestimate the fact that that when you get into the big business world, you know, you're going to see a lot of work in the next couple of years. As you said earlier, I'm making Hadoop more real time. Making Hadoop reliable. People are looking at HDFS and what other options might be because you want something that's going to have the mission critical characteristics if you're going to bet your business on it. You don't want it to be down the day you need to reprice or down the day you got a patient in front of you in healthcare. You know, you're going to see this crossover to the world where people, you know, the best thing that can happen to a new technology is that people love it and want to rely on it. But the minute they want to rely on it, it has to become dependable. It has to become more robust. It has to become scalable. It has to have all of the attributes that they're used to. Well, and you're talking about decision support before. I mean, you know, the Walmart's great examples. But in reality, there were probably only a handful of people really using those systems within organizations. They had major productivity and, oh, put the beer next to the diapers and watch what happens to sales. But there was just a handful of analysts using it. The democratization of data that you put forward means a lot more people are going to be using it in theory. And so that, in that real-time nature, really changes the infrastructure resilience that you have to have there, doesn't it? Exactly. I mean, I'm waiting to walk into a 7-Eleven and have to scan a QR code, you know, to let them know what I'm doing as a customer and what I buy from them. So they'll have the stuff that you want. Yeah, exactly, exactly. I mean, that's what you want. You want that micro-customization that you only get with the decision support. So we have some comments, tweets. Obviously, they loved your adult supervision comment, but also I'm getting a little message from some folks out there. One is, my friend's sitting with a CTO of a startup. And the question is, how does IBM's IOD, Information Fund, help the startup CTO make a rapid decision, rapid decisions faster, better? Is it human dashboards? Is it humans dashboards? I mean, how does this world relate to an entrepreneur out there who's trying to make something new happen in a big way? Well, I think for somebody like an entrepreneur, you have to figure out what the most important data is to you as you go forward. Is it your employee data? Is it your customer data? Is it your competitive data? Because if you are out there with an idea, you want feedback from your customers, because you're still trying to tune the product. I mean, in the early days of any startup, the product you go out with is frequently not the product you end up with. And the only way you cross that gap is to figure out what the customers like, what they don't like, what they're willing to pay for, what they're not willing to pay for. And I think all of those things can be gathered in a variety of ways. It doesn't have to be overly complicated. I mean, there are vendors up and down the scale, as you well know, because we've been at all their conferences where what we love about them is that they almost always start with a Xeon system regardless of whose software you run. But you don't have to pay a million dollars and sign a support and a services contract. People are trying to do stuff that you can use on your laptop. So I mean, I think that's the battle, which is... Yeah, they got two choices. They can turn key with an IBM or scale out open source with a Hadoop solution. You got Tableau, I see them here and you see all the visualization software out there. And you even have people like Microsoft, God forbid, doing stuff you can do in Excel, you know? Awesome. Or you even have guys like Oracle saying big data meet big iron. So if you want, you can bring in a million and a half dollar infrastructure to run your big data analytics, or you can run it on your laptop. So my final big question for you... I think the start up starts on the laptop in hopes they grow to the big iron phase, you know? So we're trying to do a big question for a big insightful answer in the cube here. So my big question around IBM IOD is, looking forward, knowing IBM, obviously you have a relationship with the big Intel customer, what's going on here at IOD, how do you see IBM progressing forward? As you mentioned, I see the business analytics and they have the business feed on the street and solutions. How do you see IBM taking big data today and play those next few innings out for the crowd out there in your mind's eye? Well I think even the big change between IOD this year and last year is the focus this year is really on the big data side of it as opposed to the whole portfolio. And I mean this is the show where they try and sell all their software assets but this year it's how do you incorporate big, big data solutions in the hardware, big data solutions in the teaser. And I think they're going to be looking primarily to connect their stuff together better because when you have a portfolio like that you want a way to seamlessly move the data from Hadoop to whatever tool that you want so that you can get to results more quickly and if I were IBM that's certainly what I'd be concentrating on doing. Well in the world has for the last, let's say 12, 18 months that we'll make connectors where we can connect, do your filtering in the batch and then we'll bring it over to us which is real time but you're starting to see and you'll see a lot of announcements this week and you've seen some we saw one from Adapt, you'll see some other announcements coming from MapR, Hortonworks, Cloudera that are really trying to make that a native capability. That feels like it's the right direction versus sort of a band-aid approach. Do you agree with that or? I agree with that but I also think that's going to take a while to get to maturity on those solutions. I mean quite frankly open source is moving very quickly and there are a lot of bright people doing work that I'm really, really excited about but is it yet crossed over into the bet your business side of being robust as opposed to still being something that you're experimenting with? So what's your take on, so for instance you get some sequel merging with the no sequel guys. You think people will be willing to give up the sequel practitioners willing to give up some of the sequel function in order to get that integration, that unification or will they wait? I think quite frankly, sequel's not dead and buried. It's a language that a whole lot of people know how to do stuff with and what you connect it to on the back end can change but as an API I think it has a lot of value for how you connect a lot of talent that's out there in the world to the new world of what's emerging. So I would not be surprised to see a focus on putting the sequel API into some of the Hadoop open source technology because it's just going to be a way to enable people to access stuff. I think you're right, I think that's exactly what I'm saying and I guess the premise of my premise would be that people will be willing to give up some of that traditional sequel function in sequel with that sequel API in order to get to the new stuff and then let the sequel, the new sequel guys catch up. Yeah, exactly. Pauline, thanks for coming on theCUBE again. We'd love to hear your perspective, you're a great thought leader, a great technical person and I'll see you work at Intel very relevant even though your earnings were down. I wrote a great blog post to highlight the reason why everyone doesn't understand how the PC business is moving to more processors and components. So still you guys have made some great investments, we're happy to have you on theCUBE and decision support and big data and compute powers still in demand and we're going to be covering like a blanket here inside theCUBE. We'll be right back with our next guest right after this short break.