Back in March 2015, Formula 1 team Williams Martini Racing signed a major network services and sponsorship agreement with BT, through which it is using high-performance network services across its trackside and factory-based teams.
As one of only two teams currently capable of mounting a halfway realistic challenge to runaway 2015 championship favourites Mercedes – the other being, arguably, Ferrari – Williams’ IT team, under F1 tech veteran Graeme Hackland, is under pressure to play its part.
Big slow data
However, on arriving at Williams’ headquarters at Grove in Oxfordshire, he found many of the same networking problems that Lotus had faced.
“We were not able to get data [from the circuit] to the factory in real-time, so we would schedule data transfers. For example, we couldn’t send videos for pit stop analysis so for all the Friday practice sessions we would have to do it overnight, and send feedback on the Saturday,” he explains. “Managing our limited bandwidth really limited our engineers.”
It was not just videos causing Hackland a headache: Formula 1 engineers work at the limits of their profession to implement highly technical design modifications to the various parts of a racing car, often on the fly with very little notice.
To do this they make heavy use of computer-aided design (CAD) software to design parts, often with a minimum amount of notice. Unfortunately, Williams’ engineering teams couldn’t even swap CAD files, they had to convert them to PDFs to minimise the traffic crossing the network.
Williams initially engaged with BT for a pilot project in the closing stages of the 2014 F1 season, before committing to a full partnership just before the Spanish Grand Prix in Barcelona earlier in 2015.
In the long-term, the deal will see Williams use BT’s high-performance networking services for secure and high-speed communication and collaboration, and cloud-based fixed and mobile voice services.
However, to begin with, Williams elected to use the partnership to address its connectivity challenges, and has now deployed a 100Mbps multi-protocol label switching (MPLS) network to deliver symmetric speeds between Grove and any one of the many racetracks F1 visits, from far-flung locations such as Melbourne and Singapore, to Silverstone, 40 miles up the road.
An MPLS network works by enabling data packets to be forwarded at the switching level, or layer two, instead of passing them up to the routing level, or layer three. The data packets then follow pre-defined paths to their destination, meaning the network owner can decide in advance the optimal route to get them there.
This means that service providers, such as BT, can improve the quality of service offered to the customer by defining paths that meet agreed service-level agreements on latency, jitter, packet loss and so on.
In Williams’ case, BT has further optimised the network infrastructure using acceleration capabilities, which help address latency between the team and its factory.
Williams generates between 60GB and 80GB of data per race, and this can all now be sent securely across the network to the factory, meaning engineers on location and at home can optimise the car’s performance by sharing and analysing, collaborating and making the necessary changes to improve the car’s performance, whether that is during practice sessions, qualifying, or a race.
“The 100Mbps MPLS gives us a secure link to send every piece of data back to the factory in real time. We can have the results of a pit stop analysis back inside an hour,” says Hackland.
Williams approach is similar to that taken by pro-cycling’s Team Sky to enhance the performance of riders such as Bradley Wiggins and Chris Froome, that is to say, looking for incremental gains here, there, and everywhere possible. It is from things like pit stop analysis that the network comes into its own to help eke out those incremental gains.
“We want to get our pit stops as close to 2.5 seconds as possible. Really we want to get down to 2.2 seconds, so we are constantly looking at what we could have done better, and that can be by looking at factors such as the positioning of the mechanics in the box, where the driver stops, and so on,” says Hackland.
Refreshing the network at the track has also helped Williams in other areas: it now only has to take two racks of converged server, storage and networking hardware away with it, instead of four, freeing up space in the back of the cramped pit lane garages.
Back at Grove, it will implement a 40GbE network with 10GbE links to servers and storage, and begin to move to BT’s virtualised telephone service, One Phone. According to Hackland, it was important to make sure that the transition was well-managed to fit around the constraints of running a highly mobile sports team. This will be completed later in the summer of 2015.
“This year we are pushing for the new car build to start in August, so we want to make sure that we have the network in place by then,” says Hackland. “We expect to see big benefits on next year’s car with the new network.”
Read more about networking in Formula 1
- Formula One team Infiniti Red Bull Racing uses AT&T network and UC services to stay in the running after big changes to sporting regulations
- Tata Communications joins the Mercedes AMG Petronas Formula One team to bring speedy connectivity worthy of its cars
- Formula One management, the Mercedes AMG Petronas team and Tata Communications launch connectivity and collaboration development contest