How Red Bull is racing into the future with IBM data insight tech

Spectrum Scale copes with "seriously demanding use cases" says CIO.

Red Bull Racing has a long history of leading the game when it comes to utilising top-end IT to keep pushing its efforts into the future.

But as CIO Matt Cadieux tells Computing, the secret to ongoing success is never to stop looking for ways to improve.

"We've continued to develop new products, take advantage of new cababilities to improve our working methods and IT infrastructure, as well as performance capacity and automation that the business demands of us," said Cadieux.

In fact, Cadieux now describes Red Bull Racing as "very much a data-driven company", with IBM storage products in particular now "a key part" of the firm's hugely important storage portfolio.

Cadieux describes two "high-level use cases" for data management suite IBM Spectrum Scale:

"Every year we design a new car and the car is always prototyped with tens of thousands of design changes, so all of this design work is done in the digital world, and simulated fluid dynamics and mechanic analysis to analyse different possible changes, to find the best ones," he says.

Using the technology at hand, Red Bull is now - for example - able to completely simulate crash effects on entirely digital models of cars before even a single sheet of metal has been put in place in the real world.

"We're trying to reduce weright, but can sometimes reduce the weight so much it's no longer safe," says Cadieux. Building an entirely virtualised car saves both time and money, instead of smashing up countless prototypical vehicles.

This is computational fluid dynamics (CFD), the data processing of which is run through Spectrum Scale. The team can also now fill wind tunnels with sensors on a "specialised rig" to generare even more data.

"Our entire product design process is digital," enthuses Cadieux.

If a car is being tested, it can carry up to "a few hundred sensors," reveals Cadieux.

"If the car's running in a race there's fewer sensors - usually about 100 - the amount of data we capture in a race is 100GB, that we're sending from the track to homebase in the UK."

The HQ also receives a live feed from each race, and the team farm data from broadcasters and elswhere to help "improve [Red Bull's] tactics". It's a lot of data indeed.

Race Day

Along these lines, the second use case for data management is the race track itself, on the day of a race.

"Over the course of a race weekend, the car runs for three days," says Cadieux.

"We have practice sessions on Friday and Saturday mornings, and when we qualify, it means we can no longer change the configuration of the car, and then we have a race on Sunday.

How Red Bull is racing into the future with IBM data insight tech

Spectrum Scale copes with "seriously demanding use cases" says CIO.

Throughout the race, Red Bull can gather a number of inputs from the vehicle in terms of telemetry, audio, video and other input to help optimise it, as well as their tactics.

"We believe we're leaders in this kind of activity, says Cadieux.

"But even though we're very capable today, there's still huge potential to do better," he adds.

But what made IBM's offerings stand out in a crowded data management market?

"If you go back seven or eight years ago, computationall fluid dynamics (CFD) was our first venture into high performance computing, [but with] IT using other types of infrastructure. What's really changed is we've taken some of the early lessons from HPC and we've now applied tthat to many more use cases that are [now] mainstream in the company."

In other words, Spectrum Scale has a wider reach than previous - or some current - technology, saving time and - perhaps most importantly - money.

"We're using IBM Spectrum Scale and that file system to manage CFD, but also to manage metrics and simulation across aerodynamics and vehicle dynamics across the business," says Cadieux.

"So we used to have things that ran on individual machines and then scaled up, and what we've been able to do is take those increasingly demanding use cases and put them all through the Spectrum Scale environment, which gave us much more capacity, much more performance and allowed us to high levels of storage management, to take data from expensive disc and demote it to more affordable disc," he says.

"We've been able to consolidate actiivties in the company into a shared resource controlled by IBM software-defined infrastructure with Spectrum Scale."

Cadieux describes an enablement to work in "an agile way" with Spectrum Scale.

"We found a bottleneck - constraining what engineering [did]. We've been able to scale up the Spectrum Scale envorment and stay ahead of the curve, and be very responsibe to the business as it continues to evolve in improved simulations of analytics."