How Formula One works with AWS to drive data insights


Beyond the continuous upgrades to the cars throughout the Formula One racing season, there is a constant stream of tech innovation driving fan engagement and supporting broadcasters who want to put viewers closer to the race action.

For AWS principal sports partnership manager Neil Ralph, Formula One is fundamentally an interesting use of technology, which demonstrates scalability of cloud computing services on AWS and how to use these services to solve very complex problems.

Discussing the role AWS plays in the sport, Ralph says: “We collect a huge amount of data. In fact, this generation of Formula One cars creates more than a million data points per second.”

This volume of data, he says, comes from the sensors that are on board the car along with the data collected from trackside equipment that Formula One deploys at every race circuit. Ralph says there are huge opportunities for the teams themselves to leverage the technology to improve team performance and their race strategy.

But, for Ralph, the partnership between AWS and Formula One is focused primarily on “leveraging data to unpack the complexities of the sport of Formula One for the benefit of fan experience”. This, he says, has meant that AWS has worked with Formula One to build data-driven and  artificial intelligence (AI) driven insights that are broadcast to hundreds of millions of fans worldwide.

Although from the outside, it would seem Formula One is about the race, qualifying and driver and team strategies, this data, as Ryan Kirk, lead cloud architects and team manager at Formula One, notes, is used to build stories.

“Formula One’s all about telling a story. You’ve got the main race, but there’s also so many stories that kind of happen in the background and a lot of these are data-driven,” he says. The technology Formula One has deployed from AWS has provided a way to deliver real-time insights using real-time data streams collected from million data points per second, processed in real time.

On top of this, Kirk says: “F1TV at peak will stream out at about 6 Tbps from the AWS CloudFront CDN. That is a lot of data being consumed in real time.”

Proof of concepts deployed at live events

From an IT management perspective, Kirk says: “Every race is different. New new features and new products are being released at every race. We’re always looking ahead to the next event. It’s very constant.”

While the sport has a break during the summer, it is a busy time for the teams to build out new performance improvements on the cars. For Kirk and the team at Formula One focused on the digital and fan experience, it is when they spend time developing a lot of new features.

As Kirk notes, there is also a lot of optimisation of the back-end infrastructure. “We have races during the year, so the more sensitive changes we make are deployed throughout the summer break.”

The updates tend to be tweaks or new graphics, such as AWS Insights, which was released earlier this year. This, according to AWS, uses distinct data points to inform each insight. F1 enables fans to understand how drivers make split-second decisions and how teams devise and implement race strategies in real time that impact the outcome of a race.

“It is all about expanding the product to give the fans,” says Kirk. “There are more stories to tell and we want to create a richer viewer experience for the end user, whether that is on our digital domain, or what is broadcast to your home.”

Among the challenges Formula One faces is that prototypes and proof of concepts can only truly be tested during live racing events.”Testing them during an event allows us to see what value they bring.”

Kirk says that there is a regimented testing procedure, but, as he points out: “Unfortunately, you can never simulate real life and we have to leverage race events, especially from a data perspective.”

What this means in practice is that the internal team previews the new functionality during the live race event where it is being tested. But the new functionality is not made available to broadcasters or fans.

“We can validate the data and validate the output. And then if we get all the ticks, we go forward for launch,” Kirk says.

After going live, the new functionality migrates to a maintenance phase. The project team then goes back to the start, identifying new priorities and cycling through the work to get the new functionality live.

The role of AI

AI has been among the themes has featured in the work AWS has brought to Formula One this year. AWS’s Ralph says: “Typically, we want to help remove the cognitive load from the F1 broadcast team, to allow them to do the fantastic job they do putting on a great television show.”

He says that AI is used to help broadcasters navigate the historic data repository of historic facts and statistics without having to write complex SQL database queries. Instead, he says, they are able to ask natural language questions with a chatbot that then queries the databases.

When something does go wrong, there tends to be plenty of conversations. Kirk says that details sometimes gets missed, which is where generative AI can help. “The time to resolution should significantly come down,” he says. “The tests have been extremely positive and it’s using AI, which is super cool.”

As AWS’ Ralph said earlier, the work with Formula One enables the public cloud provider to showcase its technology. “I think one of the reasons AWS has these partnerships with the likes of Formula One is because sport is a language that transcends industry sectors,” he says.

“So when we demonstrate the use of AWS technology in an environment like Formula One and hundreds of millions of people around the world watch, some may be intrigued by the use of that technology and look further into how it relates to their own businesses.”

For instance, real-time data insights are something that can be applied in many businesses. “Observing how Formula One does it can sometimes inspire companies to look at how they can do real-time data insights, leveraging high performance compute on AWS,” Ralph adds.

Download this podcast >>



Source link