 Okay, we're back inside theCUBE's SiliconANGLE.tv and SiliconANGLE.com's flagship telecast. We go out to the events and we extract a signal from the noise and we want to find the smartest, people most entertaining, people we can find to share knowledge with you and extract the signal from the noise and share that with you. I'm John Furrier, the founder of SiliconANGLE.com. I'm Joe with my coos. I'm Dave Vellante at Wikibon.org. We're here with Kim Stevenson, who's the CIO of Intel. She gave a keynote this morning just after Dave, Dave Donatelli. We're here live at HP Discover. Kim, welcome to theCUBE. Thanks. So, we're talking off camera, Northeastern Alums. Yeah. I'm having a very social morning here at HP Discover. My history is HP. Now, at Northeastern alumni, we're at the school. I actually did get a computer science degree and it's on my resume, so it's true. I'm turning on the folks who want to check that. Next job interview, I'll make sure it's on there. Well, Kim, welcome. I didn't, funny enough. I had an undergrad in accounting. So, what do you think of your keynote was fantastic? Big data, it was a big theme. Tell us about the reaction and what you thought about the keynote. Yeah, so I tried to bring to the keynote the business angle of what's happening with the amount of data being created and how that can translate into business opportunities for basically everybody with a real emphasis on moving quickly. The first movers are going to have a real advantage and it is complex, but the reality is we've got the ability to bring in all the data, synthesize what's important, and then move out. And I think the next couple of years are going to be huge learning years, but I wouldn't wait. Yeah, it's absolutely, we're big data fans and we're all over that, we love it. We're pro-big data, so the first question, I want to get your definition. From yours, everyone's got a different definition. How do you look at big data? Tech keeps looking a certain way, business people look another way. You're in kind of both sides of the fence. You're on the infrastructure side and you're also looking at the apps. I mean data's coming from applications, users, machines, systems, it's a tsunami. So what is your definition of big data for the folks out there? So I think it's basically all information created. And so whether it's machine generated or it's human generated, all of that information fits into that big data envelope. A lot of the important parts of big data though are the things, the pieces of information that we've ignored or haven't been able to contextualize in a systematic way up until now, up until now. And that's where you're going to get the real value. We've been working in IT on corporate decision making and workflow for a long time, business intelligence, but bringing in that contextualized human information is going to make it very powerful. We've had Pauline Nist on the queue, she's going to be on later today and she's going to give us probably that systems perspective. And she's- Do you know Pauline? No, I don't know. She's a character. She's phenomenal. You're a character too. But she'll give us the geeky version, but I'd like to get your perspective on something that we've been talking about at SiliconANGLE Wikibon and inside theCUBE is an observation that for the first time in history, a business can actually instrument their business end to end. Yes. How they acquire it from employees to product development, to marketing, sales, to service, everything, closed loop, full instrumentation, that's data. So as someone who's in that CIO role today, you've got to think about that. What would you share with folks kind of where we're at and if that is a premise, do you agree with it and how would you start putting that together? Yeah, so I do agree with the premise. And I sort of look at it in terms of there are 10, 15, somewhere in that range, corporate business processes. They might be ordered a cash or it might be demand generation through customer life cycle support, whatever the case may be. And that's how IT really has to think about the entire system, that end to end system and how we've been siloed over the years. But those interaction points, the seams of every handoff is where I think the value comes from in passing that information. So I'll give you our manufacturing example. We have evolved over the last 10 years to fully automated factories. And so as a wafer is moving through the factory and it completes a process step, all of that data around every die on that wafer and what happened in that previous process step is transferred to the next tool so that that information becomes the baseline information of exactly what happened, not what was supposed to happen to the next processing. That has improved our yields in manufacturing but in today's wafers, we're at nearly 750 terabytes of data on every wafer when it's end of production. And you talked about how you gave a number of examples of how Intel is using data inside of the organization and it's changing the metrics that you're actually using. You gave an example of MTBF, you said we still measure MTBF but that's not the primary decision making point anymore, is it? Right, right. We're trying to prevent, move into that whole proactive, the predictive analytics part of that data so that mean time between failures isn't the key metrics, it's when do you start to see degrading and then you repair before you ever have an incident, make it totally transparent to the business. It changes the whole game on the execution. With that in mind, I wanted to ask you about, you said big data is all information. Storage has been a hot sector mainly because it's moving from spinning disk to, you got flash, all kinds of IO and real-time analytics so it's like a perfect storm around this whole data movement, kind of transformative. But there's data warehouse and business intelligence systems out there that were kind of built the old way. And so as you talk about the premise that you guys are talking about on the keynote, is it a data, do I start with the databases? The collection is obviously important, right? If it's all data, you got to collect it. How do you get your arms around that? Yeah, I think deciding what you collect, because it can be daunting. It is overwhelming the amount of data. So making, that you have to make a decision on what you're going to collect based on what business decisions you're trying to drive. And you're looking for things that will create inflections in the business. The traditional business intelligence, et cetera, they're all structured data. And it's not that that isn't useful, it is useful. But layering on that unstructured, contextualizing that information is where you get the incremental value. And it could be, so we do stuff with supply allocation, to make revenue forecasting. So look at, I would say, look at the big corporate business processes. And where do you need an inflection point in the level of accuracy or intelligence around those decisions? And that's where you need to start. And that's hard, because a lot of times I don't know what I don't know. But you said something in your keynote, you said social and big data will completely change businesses. And then the other thing you said was that most organizations really aren't using social. You know, that much. And then you gave some examples of Intel. And your social media examples were good. They weren't earth shattering necessarily, but they were practical. And some examples of big data. Why do you think that, so big data I understand, because it's so hard, but social, you know, it was kind of in the Facebook era. You know, we got a number of IPOs and you can debate whether or not they're good or not. But I mean, it's been around a while. Why do you think that most enterprises are so slow to hop onto social? They're still skeptics. You know, so I think history always teaches us something. And if you look at brick and mortar stores at the end of the 90s into the early 2000s versus the e-commerce stores, what prompted brick and mortars to put up websites and do e-commerce? And they got disrupted. They got disrupted, right? And economics were good too. Yeah, and so, and I think that that same phenomenon that's happening with social, it's an investment upfront. So you have to justify the investment and the ROI calculations are soft. Brand awareness is going to go up. And so, it's hard for successful traditional companies to take the leap. My contention is if they don't take that leap and use the technology and innovation that's available, they will be disrupted. Yeah, and I think that's a good point. I mean, the ROI is soft because there's no real data analytics yet. Right. So what was that loop on like a click on a search engine, right? Delivers a result and that. So that's a big data opportunity. It is a big data opportunity. Yeah, especially with mobility. How does mobility affect the big data vision around from your seat? Obviously you have workers out there and you have devices at the edge. Well, and I started saying the simplest forms, it's just another data source and data fee. When you have to start analyzing the data at the point of creation though, and they're on multiple different kinds of device types, we have it into a concept called the compute continuum, where no difference between the experience delivered on whatever type of device that you happen to be on. And we would tell you we're, we would call the strategy port of choice for the operating system layer. But that compute, that continuum of experience on device consistency is a really important aspect to taking all that data in. We're here with Kim Stevenson, Intel CIO, giving a perspective on big data and just IT. My question that we've been debating and we're trying to understand this is unknown data for us in this whole real time analytics is a big, huge innovative. The iPad has shown C level executives what they could actually see and that we've heard direct anecdotal comments. CIO is going to IT, give me stuff on this. So that brings up the whole dashboard conversation, business metrics, real time analytics. So batch processing is moving to high availability and real time. That's hard, right? With all this data that you're storing, you got to have, get a move in active data and there's all kinds of tiering solutions. What's your view on that in terms of where that's coming from? Is it coming from IT or is it coming from the business line managers? We're seeing a mixed bag on that, some saying I wanted the business driver to be X, IT gets the request, or is IT enabling that kind of dashboards? Because it's still early, so I'm not sure. So I'll give you my perfect MBA answer, which is it depends. You should be a consultant. Yeah, yeah. So no, you know, I believe IT can take real leadership by helping the business to understand how you could create these dashboards. We called our social media dashboard for marketing our social cockpit, but it's a dashboard. Is it perfect? Probably not. Will it change in the next couple of years? Maybe the next six months? Probably so, as we learn more about. But waiting for that perfect answer is probably a bad idea. But I see a lot of businesses pushing their IT shop. I think the single biggest problem that IT organizations have is the velocity at which IT moves is much, much slower than the pace of business and that is something that we have to address and fix. What would you share with other IT executives and leaders out there around how to position themselves and prepare themselves for the big data movement? Like you said, it's inevitable. It's coming. Yeah, yeah. What steps do you take for the folks out there who are kicking the tires, knowing they got to jump into the big data paradigm and start rethinking the value chain of their business? Well, and maybe you could answer in the context of that last statement that you made about the misalignment of the velocity of business versus IT. Yeah, so the first thing I tell people, so cloud is your opportunity to, people talk about ROI for cloud and asset utilization and all of that's true. But the real opportunity for cloud is the velocity at which you can operate. So you'll have a provisioned infrastructure with the storage, the network, the compute available and then you're just layering on the applications. That'll cut out an enormous amount of time in what IT can deliver to the business. So I do think that there's opportunities to close that gap between the business and the way IT moves. But I think back to the data question, I would pick a few areas that the company needs better information and I would focus on bringing in that contextualized human information from non-traditional sources. People call it unstructured data and adding that context is probably the single biggest thing to get you a seat at the table. So I talk a lot about the IT labor problem. If you look at the amount that's spent on labor over the last 10 or 15 years, it's escalating. It's probably two thirds of the spend around IT and everybody talks about 70% goes toward running the business, 30% goes to managing the business. So you've got these decades of complex processes built out that have been hardened. So you remember we went through the business process redesign phase. We have to go through an IT process redesign, maybe around data, to actually attack that problem and solve that agility alignment issue. Well, some companies over the last few years they've gone through data center consolidation, apps rationalization, so if you're already there, we've made some investments to modernize our core enterprise applications, as an example. That gives us a lot of flexibility. If you haven't made those decisions, you're going to have to make those decisions. And so you're going to play catch up for a little while. In your view, do modernizing applications through the use of new tools and new practices, does that cut away at that sort of legacy process base or are we going to see similar types of calluses built up around those processes? I think it's going to cut away at it, but you have to keep looking out. Are you continuously moving forward? It isn't a one step, so it's not like a stair step. You do this once and boom, you're there. It's a continuous, we have a pretty good appetite at Intel for innovative startup firms and we bring a lot of companies in to look at their technology. Some we use, some we don't use, but part of that is cultural to build, keep pushing ourselves. And that's one thing I think is never done in an IT organization. Pat Gelsinger said in the Cube, Pat Gelsinger wrote the icon, he said that Intel marches at the cadence of Moore's Law. And so does Intel IT march at the cadence of Moore's Law and has that helped with your alignment? Yeah, it does. I mean, we not only march at the cadence of Moore's Law now but with different products and core silicon products as well as system on chip products, we now have shortened cycle times and so we're moving twice as fast and I wouldn't say that we've conquered, they keep up the pace of the business in IT yet, but I think that we've made great improvement and it's recognized by our business. That has really the point of my question. I mean, I think Intel is inherently more aligned because of that. Now, you've got other complexities, size and velocity of your business, but I think that you've compensated a lot of those just through your mission and your focus. Well, we're really excited to speak with you again. We're fans of Big Data. We built our media business from the ground up on Big Data. We actually use predictive analytics as a way to identify trend data within our target verticals to write our stories and because of that, we don't have to do any better advertising and funds are the data acquisition funds, the operation. So we're completely built in the ground up so we can tell you, for example, in our business, we can say, we can look at what's being consumed and what the target audience in our community are actually talking about in real time and it's changed our business because it reduces our risk for accuracy of stories. So one of the things that I said this morning and believe is that we're heading the place where corporate decision making is going to change the way it happens. So if I asked you a question about when you see that your readers and followers don't agree with something you print out, what decisions do you take with that data? Well, for us, we've looked at the social crowd sourcing, one we're instrumenting the crowd that we're serving. Something you talked about in your keynote. So first we have to do is instrument the crowd that we're serving. At the same time, we recognize that the crowd is part of the production process. Unlike the New York Times, which feeds them an opinion and a story, news, they don't really take them into account in the production process. So for us, we actually write stories, we'll reiterate, we'll follow up, we'll write multiple follow-ups and so we look at the crowd as a production element in our operating class. That's what they think. And so. I said we're experimenting with crowd sourcing and we're finding that I'll say the critical few are really powerful to listen to and take action based on. So we've found, our IT organization has found through these crowd sourcing applications that we're rolling out, how to isolate the true, highly skilled individuals from the herd. And that's a really powerful technique. The blended average of sentiment is, it makes your revert to the meaning. I'll share this dashboard just to give you a little taste. So we have an audience of 2.5 million unique users that we've identified as our audience. And that's now basically a panel. So in real time, we can see what's trending within our audience. And on this side is a clustered community classification. And so I can pull up at any given time, you want to talk about Hadoop, because we'll be at Hadoop Summit next week. I can pull up everyone who's in the Hadoop classification. These are people, these are real people. In our 2.5 million, this is the Hadoop sample. I randomly just picked this random guy, see what he is. You can see that these are like real people. So these are like a panel, but we're not out there trying to grab them and sell them something. This guy's a data engineer for JYR employed, employing Hadoop technologies. So he's a data guy. So we don't yet figure out how to, we don't want to go out there and just grab them and try to, but we monitor them. We harvest their conversations, look at their sentiment, figure out what's going on, and we can ping them once in a while. But again, this helps us data. So I'm a big believer that the instrumentation is hard right now. So that's why I think the database angle and this whole flash technology is interesting because the storage seems to be the bottleneck, the storage and the network. And you talk about social plus data transforming businesses. You know, we live that, so we resonated really well with us. So how about HP? I mean, why are you here? What draws you to HP Discover? So Intel and HP, we have, we're a very strategic close relationship, have for years and years and years. They're our biggest customer, but as the CIO, right, I'm here because we partner really closely with HP to help improve what we do with Intel and their infrastructure is key. It's key to my operations. So I wanted to be able to share some of the things that we've done and help them and, you know. Learn and return. So what cutting edge stuff are you deploying that you can share with the folks out there? Because that's the Intel you guys have more as law, but you also have to push the envelope and demonstrate the leading technologies. What do you have in production that's what would be quote, bleeding edge and that's exciting you right now. We have two environments that you would call cloud. One is for our office and enterprise applications and we've got 70, almost 75% of that environment virtualized and all in the cloud and we are provisioning services, new applications in that environment in under an hour. So that impacts all Intel employees because it's the office and enterprise application suite. And we had to instrument that. We've been at this since 2009, middle of 2009 and we had to instrument and automate the entire path to production and then we extended that into our DMZ. So all of our external feeds coming in are also operating in that same cloud environment, same architecture, but it's a different instance in the DMZ. Very proud of that and has really dramatically changed the way we operate. So I'll give you an example. We run an application store at Intel. It's called AppUp and we launched Angry Birds. So we knew marketing was going to launch Angry Birds. We'd worked closely with them. They expected the capacity to go up. We provisioned in our cloud the capacity to go up and however, when you give out Angry Birds for free, you get a lot more downloads than you can expect. So we came to this screeching, hauled long queues and because we were in that cloud environment, we were able to provision up that capacity. Even though it was unpredictable how much it was, we were able to provision that up. It took about 24 hours to clear the entire queue because it happened so fast and we've been learning from that experience ever since. So now we're taking that office and enterprise environment and we're bursting it out into the public cloud to create even more of that burst capacity so that that won't happen again. Our second cloud environment is really for our product development, our design engineers and that is a massive compute infrastructure, almost 50,000 servers, hung together, what I would call a grid in old parlance but basically a clustered type of cloud that allows us to improve the throughput time of every engineering job that happens in Intel. Okay, Kim, Stevenson, we've got to get in the hook. Intel CIO, thanks for sharing inside theCUBE. Great to have you. We love the conversation. We can go for another half hour. It's so exciting. Thanks so much for sharing your knowledge and your perspective. We'll be right back with our next guest after this break. Thank you.