“In our case, we practise in the morning, qualifying at midday, race at 4pm; so it’s an extremally tight schedule, and whereas [in Formula One] engineers would have six, eight hours to look at data we have 20 minutes so that creates a very serious challenge,” said Filippi.
So, time constraint and potentially unreliable network infrastructure present some tough challenges that the DS Virgin Racing team had to overcome that are absent in big data crunching in the normal enterprise world.
To do this, the team effectively created a mobile datacentre with the help of a task force from HPE.
The task force from HPE came up with a platform for hoovering up the data in the form of two powerful servers backed up by uninterruptable power supplies to ensure the data gets from the car to the engineers even if there is a power outage at the race track. One server travels to the track with the team and the other currently acts as a backup at the team’s headquarters.
“We have a very powerful, quite expensive server for the size of operations trackside and we have the exact same [setup] in England, which in the future we are going to use to create a [operations centre], more engineers working remotely supporting the trackside guys,” said Filippi.
The need for this level of infrastructure is necessary as if the team cannot get hold of the data they need, the fight for a podium finish is basically over before the race has begun.
“For us data is absolutely at the heart of everything we do,” said Filippi. “If we lose the data of a session we pretty much go into a race blind and have no chance of winning it.”
“We don’t generate that much data in terms of sheer quantity, generally a few gigabytes per session; it’s nothing crazy. What’s really tricky however is that all of this data comes in different channels,” Filippi explained.
“And really all this data only make sense if you synchronise all of it; in isolation, it doesn’t mean much.”
Enter HPE’s Vertica relational database and analytics platform. Running on the trackside server, Vertica uses the compute power on offer to suck up all the car and team data and output it into the team’s racing software in a single format.
With the system in place it takes a mere five seconds from the moment the car enters the garage to getting the data on the HPE server, then two seconds after that it is in front of the engineers eyes ready for analysis. The data is also synchronised against the time of day which makes it easier for engineers to spot the strengths and weaknesses of the driver and the car at precise moments.
And the impact of this system has certainly been felt by Filippi and his team.
“For us that was really the thing that changed completely how we worked, because that allowed us to get all of the feeds into the Vertica server and output into one pipeline that goes straight into our own proprietary racer ID software,” he said.
Once Formula E matures, with it entering its third season at present, the governing organisation of Formula Racing, the Federation Internationale de l’Automotive (FIA), will open up more data channels for race teams to crunch and analyse; currently it limits the amount of data channels on offer in order to prevent the teams with the most money from gaining an unfair advantage through exploiting very expensive data crunching platforms.
But DS Virgin Racing is already seeing positive results from its data use, with team climbing up the ranks and posting solid performances.
For HPE, the partnership with DS Virgin Racing is a good showcase of its technology and capacity to handle critical data in a challenging environment, which means enterprises looking to make use of big data analytics in more stable situations can have some faith in the tech giant’s expertise.
But automotive racing is not the only sport with such partnerships; The Americas Cup sailing racing also leans heavily on data wrangling systems, notably with the partnership between Land Rover BAR and Dell EMC.
Quiz: Are you clued up with cloud?
Page: 1 2
New York State Comptroller reportedly calls for votes against Twitter, Meta directors over failure to…