 Okay, welcome back everyone, we are live in Las Vegas for IBM Pulse, IBM's big cloud show. This is theCUBE, our flagship program. We go out to the events, extract the studio from the noise. This is SiliconANGLE's exclusive coverage of IBM Pulse. I'm John Furrier, the founder of SiliconANGLE. I'm joined by co-host Dave Vellante, co-founder of wikibon.org, research group here within theCUBE organization. And our next guest is Matt Kajou, CIO of Infinity Red Bull Racing. Welcome to theCUBE. Thank you. So Red Bull and Racing kind of go hand in hand. Let's talk about technology meets racing. Okay. As a CIO, everything's in your hands. Data-driven cars these days is all the rage. So tell us a little about your environment and the application, and then we can dive into some of the fun facts. Okay, yeah, no problem. So my team is responsible for the applications and underlying infrastructure that we use to design, make and race Formula One cars. So we have over 200 applications in the company and we make heavy use of simulations for things like computational fluid dynamics, for example, to design the shape of the car. We also do a lot with data analytics to make the right decisions during a race event. So our infrastructure, we have hundreds of servers. We have almost four petabytes of managed data. We have a very sophisticated network and we've established a private cloud that allows us to meet all the business demands for simulations and data analytics. And we've done that with the help of IBM platform computing. You go back three decades ago, it was the dawn of CAD software, sail boats to the America's Cup, Formula One racing, et cetera. And then you're good and hope for the best. Right now, fast forward, everything's data-driven from conception, design, build, and then even live and then post-mortem of say Formula One. So take us inside kind of the magnitude of the data and what goes on. I mean, honestly, real-time telemetry coming off the cars during the races. I mean, is the cars fully instrumented 100% and then what's the CIO challenge behind them? Under the hood, so to speak. So the car is instrumented. It has more than 100 sensors on it. And when the car is running, we have engineers at the track looking at telemetry but we also have an MPLS connection back to an operations room at our factory in Milton Keynes, England. And we have 20 engineers that are consuming telemetry and really understanding at an in-depth level what's happening with the car, how hard we can push it to finish as high as we can on the race. So my big challenge is to make sure all of the infrastructure just works. Engineers get the data they need in a timely fashion so they can make the right decisions. Share to Anando with the fans out there about just an example where you had the real-time data, a tweak, a change, something significant, a game-changing around the product. Yeah, I think the most compelling example we had was in 2012 when Sebastian Vettel was trying to win the drivers' championship. It went down to the very last race. And on the first lap of that last race, he was in an accident, the car was damaged. And we basically had to micromanage the car for the entire length and do just well enough in order for him to finish and win the championship. So... No, he doesn't finish. He doesn't win. He doesn't finish, Alonso wins. He's second. And going second in Formula One is going nowhere. Yeah. So, yeah, with the real-time feeds, and we were able to go and look at forces in the operations room compared against the design and what the car could withstand. And we were able to go and instruct Sebastian on how he could nurse the car. And he literally did just enough, and he won over the whole season the championship by two points. So he's got very low tolerances here, right? Very, very low tolerance. You get it. That's right, that's right. So, yeah, that was the most high-profile example of data analytics and where it's made a big difference in the race outcome. So how much is that? What would he have done differently if you didn't have the instrumentation? I mean, nursing the car, what does that mean? Like, how to take a corner, speed, is it all that? And he probably wouldn't have been aware of that? So he had to change the engine mode, for example, because the exhaust had been dented. And if he pushed the car as hard as possible, the exhaust would have cracked and the car would not have finished. So we had to instruct him on how to set the control systems on the car. We had to instruct him on how hard he could actually push it accelerating. When we took a pit stop, we also had to adjust the wing angles to adjust the balance on the car. And all that was on the advice of people back in England. Simulating. Yeah, simulating, understanding what was actually happening via telemetry and recommending back to the mechanics at the track and how to fix or set up the car and adjust. So real time, you had engineers in England, guiding you guys, translating it to the driver. And there's certain things the driver can do in real time, certain things he can't do that you had to do in the pits. That's exactly right. So how far has this come in the last 10 years? I mean, maybe you could describe sort of that journey. Yeah, well, 10 years ago it was not possible. We didn't have the data networks that could get the data back to an operations room, nor did we have the analytical tools that allowed you to digest it, simulate things, and take very quick decisions. So we've now had an operations room for about seven years. We've taken lots of baby steps, adding incremental capability. And right now we can't do without it. It's a big differentiator with track performance. So 10 years ago it was gut feel? It was gut feel. It was, yeah, experience. People at the track had telemetry. So the people at the track could see the data with tools that were not as sophisticated as what we have today. And today we can now include all the brains back at the factory, get them involved. So we have more eyeballs on the car, more specialization and expertise around engineers in the car. And they can get the data real time, no matter where the race track is in the world. The data goes back to headquarters in England. How did you deal with potential conflicts and discrepancies in opinions? Right, yeah, we say they can't take the human out of the equation, humans at the last mile. So how did you deal with that? Is it the driver's call? Is it the pit crew's call? Is it your call? No, ultimately, so the pit wall at the track is the lead decision makers for race. Christian Horner, who's our team principal, number one guy on the team, ultimately takes the calls if there's a hard decision and he'll take advice and recommendations. But ultimately, somebody may have to take a hard decision. Right. But mostly it's data-driven as much as can be. So maybe talk about the tooling a little bit. You said that was a big advancement. Obviously being able to get the data to the right place is fast enough to the pipes and infrastructure. But what kind of tooling are you talking about? What kind of advancements have we seen there? So in the infrastructure, I mean. Yeah, so what we can do is with telemetry, we can do a lot of post-processing for the math channels and we have a grid computer or a few grid computers in Milton Keynes in England. So we can run a lot of sophisticated post-processing and really get an understanding of what our car is doing, recommendations on race strategy and we can also get a very good idea of what our competitors are doing, figure out what their weaknesses are, attack and exploit their weaknesses and where they have strengths, try to counterbalance the strengths. So we run Monte Carlo analysis real-time, for example, on grid computers in England and all of this provides a decision-supported advice that gets fed back to the track in real-time. So that's awesome. All right, now, we hear a lot of talk about cloud at this show. You guys have built your own private cloud for the analytic system. I wonder if you could describe that a little bit. What does that look like? Paint a picture for us. Yeah, so we have two types of supercomputers that are based on grid computing and we've worked with IBM platform computing so that the environments are all defined by software. So we use Linux clusters and use LSF as a scheduling engine and we do a lot of simulations for computational fluid dynamics to design the shape of the car and the shape of the car actually evolves for every race. We focus new parts and new updates depending on the track that we're going to. So the big LSF-based clusters run big CFD, computational fluid dynamics simulations. The other type of cluster that we set up is a symphony cluster and we're using this to do a lot of the data analytics applications that we use real-time at the track. So it can get real-time in data telemetry, for example, do post-processing and make, for example, predictions about race strategy or recommendations on race strategy. So obviously, as you say, the track, you're customizing essentially the car for every race because of the track conditions and maybe the weather conditions as well and so forth. So you've got a lot of data that you utilize. You know, we were talking earlier about gut feel. I remember in the early 2000s, the Harvard Business Review wrote a number of articles talking about how gut feel essentially trumps data. That's changed quite a bit. It's still a big component of gut feel but part of the suggestion back then was minimize the number of KPIs because it gets too complicated. So my question to you is, is that still the strategy or are you able to optimize on a wider range of indicators or does it really just come down to a few key ones, you know, tire wear or whatever? No, there's so many variables that we're managing and we still rely, not so much on gut feel, but we still very much rely on experienced engineers but we try to give them better data so that they can make even better decisions than what they would have in the past with experience and gut feel. So you still can't take the person out of the loop. Have you seen situations where it was one of these non-intuitive things? I mean, I think about instrument flight rules when you're flying a plane. You usually got to follow your instruments, especially if you're IFR only. Have you seen situations where it was so non-intuitive and you either went with the data or you went with experience and maybe had a learning as a result of that. I think if I go back to the example with Sebastian in 2012 where he had an accident in the first lap, so the first gut instinct was coming to the pits on the very next lap and let the mechanics try to deal with the damage but the reality was is the mechanics couldn't fix the damage that was done. So it was people, especially in the opposite room in England saying, guys, there's nothing we can do about it, hold fire, keep them up. Don't bring them in because that's just going to kill you. Just buy time and let us try to figure out exactly what's going on here, but don't bring them in. So that was one where people in the opposite room were a step removed from the emotion, had the data and made a call that the act turned out to be the right decision. So you guys are, obviously, you're here at Pulse, you're an IBM customer. What's your relationship with IBM? Maybe you could talk about that a little bit. Yeah, so with IBM, we've been a longtime customer and IBM is also an innovation partner for the team. So platform computing technology, we're very lead users of the technology. We work in a very collaborative manner to define our software environment or use software to define our hardware environment for simulations. So we're both a customer and we're an innovation partner with IBM. So why don't you talk about the role of the CIO, you and your colleagues, we're always talking about things like business alignment and agility and so forth. What are the pressures that you're seeing on the CIO today? How have they changed and how is your community responding to them generally and you specifically? Okay, so Formula One is a very technology based sport and so because technology touches every part of the company from design to manufacturing to analytics at the track, everybody wants to continuously improve so the car is always improved but we also have huge pressure to improve our applications and improve the capability of our infrastructure and so that's why cloud computing really helps out because as the business demands explode, it gives us an ability to also scale up our infrastructure to meet all these incremental demands. And when we talk about cloud computing, you're essentially building a private cloud that might simulate what we've come to know as the public cloud, self-service, provisioning, you know, agility, lots of services, service catalog, is that right? I mean, where are you in that whole journey? Would you say that you've got a private cloud that provides comparable capabilities to the user base as that public cloud infrastructure as we know it? Are you a little somewhat behind, somewhat ahead in certain areas? What if you could talk about that? I think we're a leader actually. We've put a lot of effort in over many years now. We've been working with IBM and Platform Computing since 2008 and the environment went from something that was quite inflexible and wasn't meeting the needs of the business to now actually meeting the needs of the business, being very proactive and having an extensible infrastructure that allows us to react to change quite quickly. I wonder if I could play devil's advocate for a second on this topic because there may be some naysayers, I'm sure there are, right? There's always disagreements in an organization. You might say, well, yeah, maybe that's true that you can sort of replicate that but from a cost standpoint, maybe it's cheaper to do in-house actually. I'll be curious as to your thoughts there but what about it from an asset utilization perspective i.e. pay by the drink on demand type of computing? Are you able to sort of simulate those benefits? So historically, we've done everything on premise and it's really been the most cost effective and I think some of the offerings in hybrid cloud or public cloud just haven't been mature enough. So we've done everything on premise historically. Now the technology's moving on and we're starting to put our toe in the water and try to understand what the opportunities are around hybrid cloud. So from that hybrid point of view, we're very much at infancy point of view but from a private cloud point of view, we're very much leaders. So you would agree then that rental is most typically more expensive than ownership, right? Yeah, in the past when we've looked at it, it definitely was and because of who we are and how high profile we are, we can also get very attractive commercial terms to allow us to do things on premise and when we in the past have looked at doing things outside, the economics didn't work but I think now with all the investment and all the different companies that are innovating, I think the dynamics of that will change and exactly when, don't know yet but we need to watch it. Final question for you, I want to ask. Just kind of going forward, given your experience, actually Formula One, great, obviously marquee sport, it kind of mirrors the internet of things. So you think about it, you're dealing with real time, sensor data, this could be running turbines, airplane telemetry, it could be mobile phones, could be cars in the future. For the folks out there trying to understand this whole cloud, real time, big data world, what advice would you give them about what you've done and will you see it come? So I think number one, it's to work with your partners and form a vision about the art of the possible and know sort of where you want to head but then Big Bang doesn't work, it's taking lots of little steps to make continuous improvements to implement things. So I think, yeah, get the vision and then start working on it. Matt here, Kaju with Red Bull Racing Team, Infinity Red Bull Racing, CIO, Dynamic Environment, this is the future, you're living the future. Literally, Formula One, we think mirrors the future, so congratulations one on a great opportunity you've done. You're certainly the result speak for itself, snatching victory from the jaws of defeat with big data, congratulations. But more importantly, this is the future, real time using big data with computing power and a team of experts making things happen, that's the future. This is theCUBE, we are the future, we're here at IBM Pulse, we'll be right back after this short break.