Lars Baron/Getty Images
show image

Sooraj Shah

Contributing Editor

Sooraj Shah is Contributing Editor of New Statesman Tech with a focus on C-level IT leader interviews. He is also a freelance technology journalist.

How F1 is using AWS to enable more close-proximity driving and overtaking

For F1 fans, one of the most exciting aspects of the sport is seeing one car neck-and-neck with another, vying for pole position. While fans are already enthralled by the way this works today, those working at F1 wanted to be able to make this aspect of racing more exciting, by reducing the decrease in downforce generated by aerodynamics.

Currently, F1 cars suffer a loss of downforce when they are driving close to one another, meaning it’s harder for drivers to stay close to their rivals, and crucially, harder to overtake. For instance, a car running one car length behind another loses up to 40 per cent of its downforce. It means that the car slides around more and tires overheat.

“In 2021 there’s a key strategic opportunity coming for us, where we are making a raft of technical and sporting changes in F1 – so when we looked at that, and asked our fanbase what they wanted, they said closer racing, wheel-to-wheel racing, and possibly more overtaking,” F1’s chief technical engineer Rob Smedley tells NS Tech.

When one car is front of another it create a large wake structure of turbulent air behind it – which means cars are heavily affected by the cars in front of them.

“What we came up with was to redesign the shape of the F1 car in order to reduce the wake effect onto the following car and therefore we redesigned the cars by using computational fluid dynamics (CFD) which is essentially a virtual car in a virtual space,” Smedley says.

The redesign structure means that the wake structure comes off the car in front and then it disappears off into the air rather than disappearing straight onto the car behind, causing a massive loss in downforce.

The result is that with the 2021 car, the simulation indicates that it doesn’t lose 40 per cent of downforce but it loses about 7 per cent.

Using cloud computing to get there

It was taking F1 about four days to do each simulation using high-performance computing. But in order to build all of these models and execute them with an agile development process, the organisation had to look at alternatives – even though it was already using state-of-the-art technology.

“In order to be doing rapid developments, you want a parallel process with lots of different case iterations of designs and then get the results and then go back and tweak the design and get the result – so you’re constantly iterating designs to converge in what is the final design.

“However, four days was too long so we worked with AWS to use their cloud computing services and cloud HPC [high-performance computing] services and we got more than ten times the number of cores that were available with what was state-of-the-art and that’s got each simulation down to eight hours – so the output of the tool is CFD, but there’s no way we would be able to have the agile development and a converged solution by this point in the year without the AWS cloud and HPC services,” he says.

The F1 had a deadline it had to make in order to freeze the regulations, to ensure they’re ready for 2021.

“We would have arrived at the deadline with a much less iterated and refined solution [without AWS], and it’s not currently a perfect solution because we’ve made such seismic shifts but it’s a good solution, and having the speed and agility that AWS cloud computing services enabled made it possible to arrive at a much better solution,” he says.