 Live from Las Vegas, it's theCUBE. Covering Edge 2016, brought to you by IBM. Now, here are your hosts, Dave Vellante and Stu Miniman. Welcome back to Las Vegas, everybody. This is theCUBE, the worldwide leader in live tech coverage. Check out ibmgo.com. You'll see all the main tent general sessions and keynotes. Obviously, you'll see the CUBE interviews there as well and a bunch of content that's rolling in from social data. Matt Cadu is here, CUBE alum, CIO of Red Bull Racing. Matt, good to see you again. Yeah, nice to see you. Fresh off the keynote, great job this morning. Really always a pleasure. We were talking, it's been two years since we last had you on theCUBE. Yeah, that's right. And so give us the update. What's been going on in the last couple of years? We've been busy, so we just keep improving our capabilities. We rely very heavily on simulations, on analytics and we keep improving our tools to get better understanding of our curve and our tactics. But some of the new technology that's come on, we're early adopters, we take risks, and so our capabilities have grown a lot and our infrastructure has also grown a lot. So without giving away any secrets, what are some of those new technologies that have sort of changed things? I mean, when we first met, everybody was talking Hadoop and that was sort of, it was batch and everything's real time, but what are some of the texts that have come in in the last half a decade that have been interesting to you that you've been able to apply? So things that, I need to be careful here with some of the things that are secretive. So the scale of the amount of data that we collect and the capability of processing it and combining sources from multiple sources to do very sophisticated analysis. So the amount of data that we're processing is exponentially greater than what it was several years ago and having the infrastructure that has the capacity, the performance and security to manage that sheer amount of data, there's been a huge amount of effort to develop those systems. The amount of data is a proprietary secret. You guys don't share that for competitive reasons, but can you share this? You know Moore's law, right? You know the linear curve. I'm guessing your curve is non-linear. Is that a fair statement? So the shape of the curve is bending even more exponentially than it used to. Yeah, that's right. And things, so video processing, sound processing, telemetry, other sources that I can't talk about, but just the amount of input and the richness of that data and the amount of computational power needed to process it all and the amount of storage needed to house it all is huge. And cloud, a couple of years ago, it was there, maybe not as mature as it is now. Certainly IBM's cloud wasn't as mature as it is now and you're an IBM customer, but cloud is becoming an increasingly important part of what you're doing, right? Yeah, that's right. So we have a private cloud and because of the size of some of our data sets and our need to get answers very quickly, we need infrastructure to be on-premise, but then there's other things that don't have the same urgency. So what we're doing now is we're boosting or bursting out to public cloud and we're doing more and more of that. And we take it on a case-by-case basis and what's the problem we're trying to solve and then look at technically what's the sweet spot that would give us answers when we need them and financially what's the most cost-effective way. So we will have a hybrid, we have a hybrid now. We will use the trend is to use more public cloud but we will always have on-premise for some of the critical number crunching tasks. In your keynote, you talked about a spectrum of components in your stack, many of which involved an IBM Spectrum brand, but there were supercomputers in there, LSF, there was Power, Espera, which is an acquisition that IBM just recently brought in, which I think is file-related, obviously cloud. What can you tell us about your stack? Yeah, so IBM's software-defined infrastructure has a huge footprint in our state and so spectrum LSF is what we use to schedule all of our supercomputers and so all the heavyweight simulations and analysis tools are, it's managed and controlled by Spectrum LSF. The hardware that it's managing comes from multiple vendors, some on-premise, some off-premise. So that's one key product. Managing all the data on site, everything doesn't need to be on an expensive disk and so we have eight petabytes of data in the company today and finding things that are aged and demoting it to tape or to cheap disk is something that we have to do just to afford our infrastructure. So we use Spectrum Protect to do that. And then we also are using Symfony, Spectrum Symfony. So we do a lot of real-time analysis where we get telemetry and live feeds and we have dedicated clusters for that and it's Symfony that's managing that real-time environment. So IBM software is really integral to our engineering infrastructure. Do you keep all your data? I mean, we try to get rid of low-value things. So things that can be derived, we get rid of and our retention periods are usually three or four weeks. Sometimes you'll build on an analysis you've done so you keep it around for a while if you're still iterating it but then after a while you need to demote it and everything derived disappears but how you set up the experiment we tend to keep because we might need to dust it off in the future but there again you can demote it to more affordable, slower storage. So we use IBM Protect and rules in there to help delete and demote that data. Matt, can you speak to kind of the speed of innovation and how do you make sure that your infrastructure can keep up? We talk about the explosion of data, the explosive of all the centers that you have. How do you know that what you buy today is going to be okay for what you need to do in a year or two from now? Yeah, so our company is a very technical company and engineering is very, very technology savvy and so as they are trying to do more with simulations and analysis, we have a huge demand put on us to find the right solution. So how do we know that we're doing a good job? That the guys, our customers internally tell us we're doing a good job or at least they don't complain. How do we find the right solutions? We go out and we talk to the leaders in the marketplace either leaders or new up and coming companies and so we do proof of concepts, we do a lot of experiments and when we make a procurement we know that we're buying something that's fit for purpose. Let's talk about Formula One cars. So tell us more about Formula One cars. Give us the stats, like horsepower, how fast they go, give us some fun facts. Yeah, so F1 cars, it's the top of the food chain. They're the most sophisticated racing cars. The cars are lightweight, they weigh just over 700 kilograms. They have a power unit that has about 750 brake horsepower and that's provided by an internal combustion engine and also hybrid technology. So they're also instrumented with a lot of electronics so onboard computers, control systems. And then the shape of the car is designed aerodynamically and the car generates downforce that exceeds the weight of the car by about three times. So it's an upside down wing and that's what allows the car to go around corners very quickly is the downforce generated. So the cars are engineering, engineering extreme machines and a huge amount of engineering effort goes into designing them. Right, so Matt, one of the things we've looked at kind of automotive in general is the, blurring the lines between kind of the person and the machines, how do you look at that on the racing side? Yeah, so the driver in the sport, the driver needs to be in control of the car and ultimately you can give a bad driver a great car and he won't win or you can give a great driver a lousy car and he won't win. So both technology in the car and the driver, both parts need to be strong. And in the sport, we're not allowed by regulation to autonomously change settings on the car. The driver needs to be in control but if you look at the steering wheel and the number of dials and buttons on there, there are a lot of modes that the driver can put the car into. So we provide decision support to a driver that's data driven with analysis and simulation but the driver ultimately you're asking him to push the car to the outer limit and he needs to be comfortable in what he's doing. So it's a combination of driver and analytics and giving good advice to the driver. All right, so Watson can't be behind the wheel yet. No, and never will be, never will be. But how is that adjudicated? I mean, is there some kind of big brother watching? Yes, there is. So we have a digital radio and intercom system and the race authorities can hear everything we're saying. And then also when the car pulls into the garage, everything that's logged in the car, the authorities plug in and they'll know exactly what we were doing technically. And so in the sport, you push the boundaries to the extent that you can but we also need to be careful that we're compliant because we can be disqualified or fined if we're not. Is there gray area where? Yeah, always. Right. Always. So I presume you're pushing the envelope. Yeah, of course. I think if you're wimpy and you're too conservative, you won't win races. But then also you do need a basis to say what we're doing is legal. And yeah, and of course we push as hard as we can. If it's not been tested in the racing compliance court, then you get a ruling. Yeah, that's right. But yeah, the driver is really, really critical because you're asking him to absolutely push to the limit. And if he doesn't actually trust the car and when he turns the wheel, if it actually, if he doesn't believe the car will be in control, then he won't push as hard. So it's really important that he's comfortable with the car. And then also we show the driver a lot of the analysis and we explain what's happening under the covers so that the more he understands the car, the harder he will also push it. And then his feedback also gets fed back to engineering and it helps us to improve the car for future races. And we're talking top speeds of 200 miles per hour, that's right. And zero to, I guess zero to 60 is irrelevant, but. Zero to 100 and back to zero is about 4.8 seconds. Zero to 100, back to zero. Back to zero, yes. In 4.8 seconds. That's right. That's a lot of cheese. Yeah, and then G-forces can exceed also six Gs. So high speed corner with braking, the car and the driver have to withstand a lot of G-forces. So what do you make, I mean as an observer, I mean you can't use autonomous technology in your sport, but you guys are in the vehicle business. What do you make of what's happening with this whole autonomous vehicle, just this personal opinion? I like driving personally and really don't want to be driven, but I think there's a lot of people that commute and it's wasted time and I can see autonomous cars for many people being something that helps them and gives them more time. My personal opinion is though, I grew up in Detroit and I like driving that and still like driving. Such an interesting topic though because many of us learn how to drive in a stick shift, which is more fun. When you turn 16 in the United States, it's part of your right of passage. One of the first things you do is go get your license and you wonder socially what the implications of autonomous vehicles are. Yes, there's all these great things potentially anyway, but in terms of the social implication to a teenager and an individual who loves to drive, what do you think's going to happen there? I mean, maybe it's a hybrid world. I think that the transition will be challenging when you have people driving and making human error and you have machines talking to one another, but when they both have to coexist, that will be adventurous. Driving's fun. So it will happen with all the automakers investing huge money and with technology advancing, it will happen. So you called your vehicle an evolving prototype. What does that mean? Yeah, so it changes every race. And so we don't just design one car at the start of the season and just race it continuously. We actually tear the car apart and we rebuild it to a new specification that's targeted for the race track that we're going to. So every race track has a different shape, different surface, different weather conditions and we micromanage the spec of the car to be optimized for that race. So we design new parts, new assemblies that are targeted for specific races. And so every race, hundreds, sometimes thousands of parts, new parts get put onto it. And last year we had more than 30,000 engineering changes throughout the course of the year. So it is really a prototype. We don't mass produce. We have two cars that compete in the race. We have a few spare chassis that we introduce in the middle of the year or if we have an accident. But yeah, we, with those cars, we tear them apart and modify them for absolutely every race. So is the strategy to be a decathlete or a specialist or both? If you know what I mean by that. Is it horses for courses? Some horses do well on tight turns, others do well on the wide sweeping turns. Some cars must do well with tire wear. So... It's all kinds of trade-offs. So what's Red Bull? What is your sort of sweet spot? What do you even know? So high speed or tracks with a lot of corners, high speed or slow speed corners are a sweet spot. We don't have the most powerful engine on the grid today. So tracks that are long straights are our worst tracks. Ones that have more curves are the sweet spot. And when we set up the car, we make a lot of engineering trade-offs. And there's a lot of actually hard choices to be made. And this is also where we do a lot with simulations and a lot with the sensors on the car and compared it to simulations. And on a Friday and Saturday, when we have practice sessions before a race, we're making data-driven decisions around what parts go on, what settings go on to the car in order to set it up so that it's set up to then perform well on the race on Sunday. So all of that is very, very IT intensive, generates a huge amount of data. And is that choice cultural? In other words, you know, like, I like skiing the bumps. You know, I like going the big wide sweeping turns. It's like, you know, the tight turns is made is it because it's more fun and that's the culture of Red Bull or it's just sort of what the engineering is good at? Ultimately, our goal is to win the race. No matter what. No matter, yeah. And yeah, within the rules to win the race. And so to make the engineering choices so that you have a car that will be optimized for the race and in part of it is factoring what do the drivers like, but then also it's understanding the truck and the car. And when you have hundreds of choices around the setup of it, it's making the right ones. So it's very, very data-driven. So how are you guys doing? We're doing all right. We're in second place in the league table today. Mercedes is the top team and we're closing the gap to them and we're above Ferrari. That's pretty good. In between Mercedes and Ferrari, that's good. So yeah, we're improving and we're in a strong position. And second, next year is interesting where there's a major rule change. So right now we're also splitting engineering resources and putting a lot of effort into next year, which has some very, very big changes to the architecture of the car. And those rules, I mean, sure it's changed. It's like small rules every year, but you're talking about some big rule changes. Yeah, it's not evolution. This is more revolution. So it goes back to the big fat tires. The aerodynamic limits and constraints are very, very different. So next year's series will be very different than what it is this year. And then Renault, our engine partner, also has a lot of improvements in the pipeline. And so with our chance to innovate on the chassis side next year, where we've always been strong and with improvements on the engine side, we think next year is going to be the chance to really give Mercedes a run for the money. Fun. Oh, that's got an engineering playground. That's awesome. It's nonstop, engineering, design a faster car, both for this season and the upcoming season. But the IT infrastructure and applications needed to support it. If we didn't have robust, highly capable solutions, we could not design a fast car. Well, Matt, congratulations on the success that you're having. Last thoughts on Edge, things you're learning here or sharing and things that are exciting you? Yeah, it's a chance for me to learn. And I get to meet other people and ask them how they're addressing challenges. And I also get to meet IBM experts and some of their partners and ask about what's in the pipeline and other things that could potentially benefit us as we continue to improve. So it's a chance to learn and really enjoy the session. Excellent. Well, great to see you again. All right, nice to see you. Reving our way through IBM Edge. Keep it right there, everybody. Stu and I will be back with our next guest. This is theCUBE, we're live from IBM Edge. Right back.