McLaren Racing uses Splunk to analyse 1.5 TB of data every race weekend

McLaren Racing was one of the first Formula One teams to put sensors on its cars. Over 30 years ago, the car had 24 sensors on it to help the teams to get a better picture of what was going on during a race.

Fast forward to 2020, and McLaren Racing now has nearly 300 sensors on its car every race weekend. The company collects about a terabyte and a half of data each race weekend. That could be information such as lap times, or it could be data to help understand the load on the front suspension.

Ed Green, principal digital architect at McLaren Racing said:

The reality is that it's too much for individuals to monitor every single data point that comes in. We'd like to - because it's the way we coach the drivers, develop the car, and how we build our strategy as it impacts every part of racing for us, but it's impossible.

Over time, McLaren Racing has built its own software to handle the data - with machine learning models and systems that can analyse anomalies, to help detect any issues. One of the key technology partners that help with this has been Splunk. Green said:

What we've been able to achieve with Splunk has been a much more rapid acceleration of some of these projects and the ability to bring the data into one place. Splunk now sits across our travelling trackside rig, we take compute with us right to the edge, so all of that one and a half terabytes is collected and gets pushed backed to the McLaren Technology Centre. This means we can have a group of engineers in the UK while the race is taking place, and we can bring the data back to the headquarters, as well as analysing it at the edge on a bunch of servers or on custom software.

McLaren Racing started working with Splunk a couple of years ago to evaluate performance data - but it has now entered a new partnership with the company, which will mean deploying Splunk across McLaren Racing’s infrastructure. Green said:

This will allow us to provide better operations, more reliability and consistency. It's quite hard to describe how critical our infrastructure is - if we can't see the car [on our infrastructure], we can't go out on track, so if a server or database is not available or not reporting in the right way, it would mean not being able to compete in quite a public forum.

As McClaren Racing has always worked with teams under pressure, and collaborated globally and remotely, the pandemic hasn't meant any new challenges. However, as McClaren Racing operates at speed with minimal SLA times, it's essential that its IT operators and engineers can get a picture of the data quickly, and this is what Splunk has helped the company to achieve.

Splunk CEO Doug Merritt told diginomica:

We were really excited when they approached us two years ago - they studied the entire landscape and they knew what was coming in that you couldn't just outspend everyone else in F1 and win every race and they viewed data as being absolutely critical to all teams' success.

What we're doing is streaming processing with them…you're not going to get that real-time stream while they're racing but as soon as they take their chip and push it, you've got to be able to really make sense of that data so you can make adjustments.

With the intense workload of moving a data centre every two weeks, or in some cases every weekend, Splunk has made the task simpler of spotting trends and getting smarter alerts to help McLaren Racing’s team with decision making. Ultimately this means optimising the cars and the team's performance. Green said:

It means you can give the team a realistic expectation of how long something will take if there is an issue or you can start to predict [an issue] and work with people on the run.

Why Splunk?

Cloud computing is an important aspect for McLaren Racing in being able to scale up on demand. Splunk is making its own transition to becoming a cloud company - but Green said that his team didn't consider looking at cloud alternatives such as Sumo Logic. He said:

We were so impressed with what we were able to do with Splunk the first time around a few years ago and when we got the Splunk team in again it was clear to see advantages in the speed of data and the metrics we were able to get through.

We were able to look at some of those data pipelines coming in; it could be 100khz a second of data coming in which is an enormous about of data, so being able to look at that as a metric and bring it in and spot the trends was key.

One of the areas where Splunk is being deployed is with McLaren Racing’s e-sports team. Green explained:

It shows how quickly we can interpret data. The information coming from the games and the simulation world is almost richer - with as much data but more consistency than actual F1 cars, so that's been exciting to look at, model out and take inspiration from the virtual world and start to apply it to the physical world. It meant that in periods where we weren't racing we were still able to learn and work with Splunk.