Season 2: Episode #9

How AWS Speeds up Formula 1

Formula 1 races have always been intense, with drivers reaching speeds over 200 mph on the track. But fans wanted to see closer racing and more overtaking – so F1 Lead Cloud Architect Ryan Kirk had a challenge ahead of him in designing a new F1 car. Hear how he and the F1 cloud team used the supercomputing power of AWS to overcome massive aerodynamic forces to speed up the design process and the race itself.

Ryan Kirk


Ryan Kirk

Engineering manager, cloud and DevOps lead architect for Formula One

Read Bio
Ryan Kirk

Ryan Kirk

Engineering manager, cloud and DevOps lead architect for Formula One


Hilary Doyle: Hi, Rahul.

Rahul Subramaniam: Hey, Hilary.

Hilary Doyle: So today we’re talking about AWS as athlete, and this caught me by surprise.

Rahul Subramaniam: That’s absolutely right. I’ve sent you a link. It’s a video from the recent F1 race. I really want to hear what you think about it.

Hilary Doyle: Okay. You will learn during this episode that prior to preparing for today, my full knowledge of car racing was basically Louis Hamilton and Hot Wheels, not actually in that order. Yeah, let’s check out the Canadian Grand Prix for, in my case, the first time.


Rahul Subramaniam: Yes.

Hilary Doyle: One, two, three, click. Oh, now I’m in the cockpit with him.

Rahul Subramaniam: Yeah.

Hilary Doyle: Oh my good Lord, there are no roofs. There’s no roof on this thing. It’s like a kitted out go-kart. Oh, my gosh.

Rahul Subramaniam: Look at that cockpit view and how close those cars are right now.

Hilary Doyle: Oh, Louis Hamilton coming up from behind. Oh, he is extremely close. That is not safe, guys. That is not safe. Keep some distance between you. You’re in a moving vehicle with no doors.

Rahul Subramaniam: And that’s the overtake.

Hilary Doyle: I’m sorry, I didn’t even see the overtake. I was just concerned about them living.

Rahul Subramaniam: So Hilary, how do you feel?

Hilary Doyle: I’m exhausted and I feel like a parrot. Why do they drive so close to one another? It is not safe.

Rahul Subramaniam: Now even though it might seem ill-advised, the reality is that the F1 cars are designed with absolute precision with tons and tons of data points that have gone into understanding how to push the boundaries of performance on those cars. And at the same time, balancing aspects of safety.

Hilary Doyle: I mean, sorry, every time you say safety in the context of this word, it makes me laugh. Because safety on the road for me looks slightly different. But I take your point.

Rahul Subramaniam: The redesign of the F1 car required so much iterating that the team at F1 had to bring in some outside help.

Hilary Doyle: A pit crew by the name of AWS.

Rahul Subramaniam: Exactly. So today I’m really excited to get into how the AWS cloud helped deal with literal clouds of messy air coming off those F1 supercars. All in pursuit of closer and closer racing.

Hilary Doyle: Clouds of messy air was the name of my middle school boyfriend’s grunge band. But I digress.

Buckle up my friends. Here we go. This is AWS Insiders, an original podcast from CloudFix bringing you what you need to know about AWS through the people and the companies that know it best. CloudFix is the nonstop, automated way to find and fix AWS recommended savings opportunities. 

I’m Hilary Doyle, I’m the co-founder of Wealthie Works Daily.

Rahul Subramaniam: And I’m Rahul Subramaniam. I’m the founder and CEO of CloudFix.

Hilary Doyle: I mean, as anyone can imagine, it is taking g-forces of willpower, Rahul, to avoid the very obvious puns that could be littered frankly throughout this episode. So I will just say in earnest that we’re going to race through the beginning of this episode because we’re so excited to speak with our special guest.

Rahul Subramaniam: Exactly. We’ve got an inside track with F1’s very own cloud lead architect in a moment.

Hilary Doyle: Okay. Well, we gear up for that, quick pit stop for your AWS news headlines. That’s actually what it is. It’s a pit stop. That doesn’t count as a pun.

First up, AWS is raising its prices, or it’s actually creating a price. As of February 1st, 2024, AWS will introduce a new charge for public IPv4 addresses 0.005 cents per IP hour for all public IPv4 addresses, whether they’re attached to a service or not.
That’s a lot of zeros behind that decimal, but any new charge makes me suspicious. Rahul, what’s going on here?

Rahul Subramaniam: So the IPv4 addresses have been around since the beginning of the Internet. Now, no one anticipated what the internet would eventually become. A couple of decades ago a new addressing scheme called IPv6 was created to significantly increase the number of computers that could exist on the internet.

Now, we thought that we would run out of these IP addresses about 20 years ago. But thankfully the creation of the NAT Gateway pushed out that scenario about now. Now AWS is hoping that by charging for these public IPv4 addresses, they will be able to impact the move to IPv6.

Hilary Doyle: Smooth incentivizing. Sticking with financials, AWS has unveiled the cloud financial framework. It’s a collection of actionable guidelines that aim to help AWS’ customers optimize their workloads across different AWS services. Caveat. In order to get the most out of these guidelines, you should have a 200-level knowledge of AWS.

Let’s be honest, Rahul, the only question I have for you here is what level of knowledge are you?

Rahul Subramaniam: I don’t know, I’m not certified. But let’s see what 15 years of doing cost optimization qualifies as. I’ll check with AWS.

But I’m really, really excited about this one. AWS has finally shared a framework for cloud cost management that is pragmatic and prescriptive. I would urge everyone to check out the playbook section of the cloud financial management for each service. Make sure that you are following those best practices.

If you want automated implementation, then there’s always CloudFix to help you with that. I’m proud to say that CloudFix covers all of the playbooks and more that are currently published over there.

Hilary Doyle: CloudFix, that cheat sheet for cloud savings. That is not our last news item about guidance from AWS. They’re so helpful, especially when it comes to customers accelerating large-scale migrations.

My friends, meet prescriptive guidance. Providing time tested strategies, guides, and patterns to help accelerate your cloud migration, modernization, and optimization projects. This sounds helpful.

Rahul Subramaniam: Well, the key word here is prescriptive. Now, you’ve heard me complain in the past that AWS hasn’t been opinionated. And now finally they’ve published a huge catalog of well-architected patterns that customers can use to build their cloud native solutions.

Hilary Doyle: AWS grows its first opinion. That’s it for your news headlines. Let’s head back to the track, get into the driver’s seat, rev up those engines.

Rahul Subramaniam: All right, Hilary. It’s time.

Hilary Doyle: We are joined by Ryan Kirk, the engineering manager as well as the cloud and dev ops lead architect for Formula One. Welcome to the show.

Ryan Kirk: Thank you very much.

Hilary Doyle: We have a lot to cover, so we’ll start at the beginning. AWS really helped to co-create the F1 car we saw in the 2022 season, the standard for all cars on the F1 track. Would you please share some of the background on how this F1 AWS partnership came to be and why?

Ryan Kirk: So for many years, the aerodynamic of Formula One cars, it remained fairly static. I’m sure the aerodynamicists who I work with would be shouting at me silently for saying that.

But one of the problems that we were running into, or specifically the teams were running into, was overtaking. That was a bit of a lack of overtaking within the sport, which was noticed within. Certainly within the fans as well. That was when as a car, if you imagine two Formula One cars driving side by side, you’ve got the front one which is paving the way. Which is getting all of this nice calm air that was cutting through it, going over the car.

But unfortunately as the car behind, all this jumbled air was then hitting the car behind. Which made the aerodynamics of that car incredibly hard to overtake and losing a lot of downforce. Now, what we wanted to solve is that exact problem. We wanted more overtake and more excitement within those sports, to have those consistent battles and those dog fights.

Hilary Doyle: Close racing has become a phenomenon in F1 in the last number of years. Why is that?

Ryan Kirk: So I think, I mean taking nothing away from Mercedes, but they dominated for many years. I think that although the fans still love F1, I think they wanted to see a little bit more competition. They wanted to see those kind of middle ground teams and the lower end teams and those underdogs come through the ranks. Not be limited and really fighting for those places.

So I think that that became the priority to bring that nail-biting excitement, rather than just seeing one team just win it race after race, season after season. So we were aware that on social media and stuff that it was quite evident what the fans wanted. It was closer racing and more overtaking. So this basically laid on the aerodynamics of the car and redesigning or coming up with a specification of car that would reduce this what we call messy air, which is the air that comes off the front car.

Now, over the course of this project, it was a multi-year project. It was very iterative as you can imagine. We ended up running seven and a half thousand simulations. So put there in perspective. If you ran that same amount of simulations on a modern day high spec laptop, it would take around 500 years to achieve that.

Rahul Subramaniam: Wow. So how much data are we talking about that was actually involved in this kind of compute?

Ryan Kirk: So where we would run different simulations, so we would have a simulation of, say, it could be a wing. Changing the wing tip single digit degrees and seeing what the result was. If it gained you a 0.5% less messy air, then that was successful. So not that much data. Or it could be scaled up to two cars following each other, and we would simulate what that airflow is off two cars.

Now if we’re talking the two car element, we’re looking at probably half a billion data points. And then over the course of the project, we generated about half a petabyte worth of aero data that we were then able to play back and analyze. So from a data point perspective there were huge amount of parameters, huge amount of data points. Consequently, a lot of data that came off at the back of that.

With something like AWS and the partnership, we had that kind of infinite scalability. It allowed us and our own aerodynamicist to create a virtual wind tunnel within the cloud and scale as much as we want. Yeah, that partnership with AWS certainly helped provide that platform and that scalability to run these simulations around the clock.

Hilary Doyle: We know that AWS was instrumental in shortening the periods of time for those simulations, but can you give us an understanding of scale here?

Ryan Kirk: Yeah, sure. So let’s give an example of the historic process that we did before we partnered with AWS was we had a technical partner who would run our simulations on our behalf. This was on an on-prem data center. Now for a single simulation, we were looking at probably three to four days of a turnaround for a single simulation. When we partnered with AWS, we were able to bring that down to 12 hours for a single simulation.

That was groundbreaking to us from an infrastructure and cloud perspective because that meant that we were able to run two simulations per day. So we could run one overnight and we could run one during the day.

We then looked and go, wow, we can do this within the cloud. How can we expand it? So then we expanded into multi region, so then we would have multiple ones run running all the time. So it became a bit of a machine, really.

That process was incredibly simple. Our aero team would make some tweaks to the design, they would load the design into the system, and then the infrastructure just used to take that in and scale out as much as it needed. Using AWS’s scalability, it would scale out. And then when it was done, it all scaled back in again.

So it was incredibly easy for us, which was great.

Hilary Doyle: So how significant is data in driving the iterations that F1 or F1 teams are running on the cars themselves?

Ryan Kirk: I think it’s very, very significant. I mean, the people say that Formula One is half a sport and half a technical spectacle.

In some ways they are correct because we are incredibly data-driven. The teams are incredibly data-driven. There’s just so much data that does come off the cars that are analyzed in real time, and also after the races, that I think the majority of those decisions are very much driven by data. Whereas not that long ago really it was driven by the engineers. They were watching the car and the driver experience and then having to come up with a solution based off of what happened at the time.

Whereas now the data that we’re able to consume and record back, we can and the teams as well can make a lot of those decisions based off of that.

Hilary Doyle: It’s hard to have these conversations without hearing the letters DRS come up over and over again. Would you explain to our listeners just what DRS is and why it matters to F1?

Ryan Kirk: So DRS stands for drag reduction system. Essentially at a high level, you may see on the, spoiler, at the back, there’s a flap that sort of drops down.

Now when a car has DRS, that flap basically changes its state. It will give the driver a slightly, ever so slightly, aerodynamic edge. Now, I say slightly. It’s not slight in the race, it’s actually a huge advantage. The car is then able to pick up on a strait, you’re talking maybe single digit miles per hour down the strait. But it really is enough to get them into those overtaking positions. So it’s incredibly advantageous.

Hilary Doyle: So at a certain point after multiple iterations, simulations, adjustments, you had an updated design to put on the actual road. Tell us about that experience.

Ryan Kirk: Yeah, that specification came out. It got signed off, and we gave it to the teams. There was some nail-biting moments early on that season just to make sure that it was what we had calculated it would be. Fortunately, it delivered the result.

Hilary Doyle: I’m just going to pause our interview with Ryan here for a second. Because even though we’re getting you under the hood of Formula One’s new car, Rahul, you host another show that also gets people on the inside lane at AWS.

Rahul Subramaniam: That’s absolutely right. Every week I’m joined by my colleague and fellow AWS enthusiast, Stephen Barr, for a livestream. Breaking down the latest AWS news, sharing insights, not just our own but also those of some amazing guests from AWS.

Hilary Doyle: Bring your edge-of-the-seat questions, get them answered live on AWS Made Easy. It really does make things easier. You can find out more at

In the meantime, we’ll keep this insider chat going. We’ll get back to our conversation with Ryan Kirk.

Ryan, take us back to the pre-season. I mean, you put a new car on the track for the 2022 season. Suddenly, drivers are adjusting to a completely revolutionized vehicle. What were those first days for you? What were you experiencing personally, and what were you seeing on the track?

Ryan Kirk: So I mean seeing it there in physical form is quite something. We had developed these virtual wind tunnels in the cloud. You see these models and stuff, so that was cool seeing that for sure.

Hilary Doyle: What was the first race that you watched after you had redesigned this car?

Ryan Kirk: So the first race I believe was Australia. This was the Grand Prix, so I think it was on the first lap. I mean, we always get great racing on the first lap because all the cars are so bunched together. And then once a few cars were able to put away, and then I think there was that kind of overtaking going on. It was dotted all around the track. I think that was when we were like, okay, this is working.

The feedback on our social media platforms as well, it was very positive. So that’s when we know that it worked, I think was… Because we can say, oh, that looks great. There was two more overtakes than there was in the last race. But I think it was where you had this dog fighting. It was very evident that the cars were able to have these kind of multi-overtaking type battles, which the fans loved.

Rahul Subramaniam: Can you tell us a little bit about the services that you used under the hood with all this compute that you had at your disposal?

Ryan Kirk: So we relied a lot on EC2, the scalability of EC2. We were able to leverage AWS’s own chip set, that Graviton instances, which were fantastic as well. These were game changer things because AWS develops them. They’re more sustainable, which is a big driver for us than the standard chip sets.

Rahul Subramaniam: Have you already tried out the C7gs?

Ryan Kirk: Yeah, yeah. Big fans of Graviton. I’m sure my team, they come with some great systems, but they usually do sort of bug them. If it’s not Graviton, I’ll try and press and get them to go on those because I think they’re great.

Rahul Subramaniam: I couldn’t agree more. But was your choice of sticking with EC2 instances versus, say, trying something like Graviton, driven largely by the fact that it is the tools that you were using to run these simulations? Or were there other decisions that went into that?

Ryan Kirk: Yeah, so there was a few decisions. We knew that we needed raw compute. As much as we would’ve loved to explore serverless technologies, that unfortunately just weren’t viable for us. So it very quickly landed us to EC2.

Now, the exploration of instance types was quite a journey. Because although we wanted to get the most powerful instance to do it the fastest, we also wanted to do it efficiently from a financial aspects. So it was a lot of balancing between how do we really get the most out of these instances and that best price performance that we possibly can.

And so there was a lot of baselining. There was a lot of testing with varying instances. As well as the types of simulations that we had as well catered towards different instance types, which was quite an interesting discovery. Because stuff like Graviton would processor type simulation more efficiently than something like standard x86 type architecture, like an intel architecture.

So that was quite fun, I must admit, doing the baselining and the performance engineering of it as well.

Rahul Subramaniam: So Ryan, it seems like the cloud was the big enabler for a lot of this change to happen where you actually ran a lot of simulations yet maintaining all the safety of the cars. F1 picked AWS as the partner to do all of this.

One, what is the decision-making process to move to the cloud versus trying to do something in an on-prem data center?

And second, why AWS?

Ryan Kirk: So the first one I’d say, I mean as a company we were quite late to the party with cloud technology. We had an incredibly closed system for many, many years. It really was a secret sauce. Under the hood, it was a 200 ton plus data center that we used to chip to every single track.

Rahul Subramaniam: It’s interesting that you measured the data center in weight rather than…

Ryan Kirk: Yes.

Hilary Doyle: Yes.

Rahul Subramaniam: Dead weight.

Ryan Kirk: Yeah, exactly. So we would send this thing and in my past life I was a trackside systems engineer, so this was very much was my bag. It went from place to place, and it was our crown jewel. That’s how we got the race out.

Now when we decided to make that strategic decision to move to the cloud, it was a huge thing for Formula One because we’re essentially saying we’re coming out of this kind of closed system. We’re going to start using someone else’s infrastructure. It was huge enabler for us having that kind of almost infinite level of compute and services and global presence and distribution.

Why we chose AWS? I mean, they’re Gartner quadrant leaders. They’re especially focused on innovation, that it just made sense for us to pair with them as they’re enthusiastic about innovation and as we are. So it made for a great partnership.

Rahul Subramaniam: Ryan, moving to the cloud is a big paradigm shift. It requires a different way of thinking about things. We invariably talk about a lot of the wins or the successes in the cloud. But that win comes on the back of a bunch of different learnings. So I’d love for you to share one big aha moment for you in your cloud journey.

Ryan Kirk: Yeah, so I think that one of the things that we saw as a business as a real kind of that aha moment, that real golden opportunity, is global distribution as a platform.

So as you’d expect, we do a lot of the race. But we also do a lot for other services. Our data goes everywhere. Now, having a platform where we can distribute video, we can distribute data on a global scale to an infinite number of clients was a huge, well, this is great. We can really expand on this. We’re not limited at all by presence, infrastructure, and that side of things.

So I think that was quite a big moment in our remote production efforts. So because that is where we’re, as a business, driving towards remote production. So that was a great moment for us when we realized that.

And in terms of lessons learned, so we were quite calculated in what we moved into the cloud. We didn’t move stuff that didn’t make sense. So we’re incredibly latency sensitive as an organization. So when there’s a race on, milliseconds do count. Especially now we do remote production. If we’re based in Japan and then we’re doing the production out of the UK, then those milliseconds come even more precious.

And so I’d say a piece of advice to people is don’t just rush into the cloud with everything. There has to be a use case. I think that if you do sit down and go, what advantages are we going to have here by putting it in the clouds, you won’t have that moment where you have to reverse engineer something. Put it out and have that moment of, actually this probably wasn’t a good idea. That would be a piece of advice I guess for others.

Rahul Subramaniam: Is there any other takeaway you’ve had over the last few years that might be useful to others who are just starting out in the cloud?

Ryan Kirk: I would say even though if you develop something or move something into AWS, I wouldn’t necessarily just leave it there. Especially if it needs to be scalable, performant, AWS has a million and one ways to do things. You really can get a lot of performance out your system.

So continue to innovate, just continue to improve it and leverage the services, those great services that AWS provide, would be my advice there.

Rahul Subramaniam: Ryan, what’s your prediction? Are we going to cross the 400 kilometers an hour goal over the next two or three years with all the work that you guys are putting in?

Ryan Kirk: I sure like to.

Hilary Doyle: It’s your man on the moon moment.

Ryan Kirk: I sure like to.

Hilary Doyle: Thank you so much for making time to bring us all up to speed on F1. It’s been wonderful to have you.

Ryan Kirk: Thanks for having me on.

Rahul Subramaniam: Yeah, thank you so much, Ryan. This has been an absolutely fascinating conversation.

Ryan Kirk: You’re very welcome. Thank you very much.

Rahul Subramaniam: So Hilary, will you be tuning in for the next Grand Prix?

Hilary Doyle: I will be watching from a very safe distance, like my couch.

But I’m curious, Rahul, after speaking with Ryan, it is an epic example of AWS technology and compute at work. But what can other AWS customers who aren’t moving at 200 miles an hour learn from this F1 example?

Rahul Subramaniam: Okay, so the funny thing is as I think through our conversation with Ryan, I think it is the manufacturing industry that stands to learn the most and replicate what F1 did.

I know it sounds weird, but hear me through. We are currently living in this fourth industrial revolution. One of the core underlying driver is the concept of a digital twin. If you had all the data that every machine in your factory produced and you could create a virtual simulation of the machine and maybe even predict when the machine is not performing optimally, and of course take preventive actions as a result of that, being able to simulate and predict that is orders of magnitude more cost-effective than having a real machine actually break down.

When you look at services like AWS IoT SiteWise and AWS IoT TwinMaker, they’re already available and helping customers do just that.

Hilary Doyle: Twins, they’re so freaky.

Okay, please share your AWS stories and questions with us at Some of my very close friends are twins.

Rahul Subramaniam: And please leave us a review. Don’t forget to follow the show to get the new episodes as soon as they’re released.

Hilary Doyle: A WS Insiders is brought to you by CloudFix. They are an AWS cost optimization tool, and you can learn more about them at

Rahul Subramaniam: Thanks for listening. Bye-bye.

Meet your hosts

Rahul Subramaniam

Rahul Subramaniam


Rahul is the Founder and CEO of CloudFix. Over the course of his career, Rahul has acquired and transformed 140+ software products in the last 13 years. More recently, he has launched revolutionary products such as CloudFix and DevFlows, which transform how users build, manage, and optimize in the public cloud.

Hilary Doyle

Hilary Doyle


Hilary Doyle is the co-founder of Wealthie Works Daily, an investment platform and financial literacy-based media company for kids and families launching in 2022/23. She is a former print journalist, business broadcaster, and television writer and series developer working with CBC, BNN, CTV, CTV NewsChannel, CBC Radio, W Network, Sportsnet, TVA, and ESPN. Hilary is also a former Second City actor, and founder of CANADA’S CAMPFIRE, a national storytelling initiative.

Rahul Subramaniam

Rahul Subramaniam


Rahul is the Founder and CEO of CloudFix. Over the course of his career, Rahul has acquired and transformed 140+ software products in the last 13 years. More recently, he has launched revolutionary products such as CloudFix and DevFlows, which transform how users build, manage, and optimize in the public cloud.

Hilary Doyle

Hilary Doyle


Hilary Doyle is the co-founder of Wealthie Works Daily, an investment platform and financial literacy-based media company for kids and families launching in 2022/23. She is a former print journalist, business broadcaster, and television writer and series developer working with CBC, BNN, CTV, CTV NewsChannel, CBC Radio, W Network, Sportsnet, TVA, and ESPN. Hilary is also a former Second City actor, and founder of CANADA’S CAMPFIRE, a national storytelling initiative.