Williams Martini Uses BT in Winning Formula

London -- Broadband World Forum 2016 -- Formula One Grand Prix (F1 GP) racing is a brutally competitive sport, and for smaller teams such as Williams Martini, keeping up with new technology and car performance is a challenge.

According to Williams CIO Graeme Hackland, larger F1 teams such as Ferrari and Mercedes have two to three times the staff and technology budget that his organization does, but they still have to keep up and be competitive. Speaking at the Broadband World Forum in London this week, he laid out the evolution of IT in F1 GP, starting with 1979, when it took 20 minutes to download data from one lap. He described 1989 as a watershed year for F1 racing, when they transitioned from a 64KB "black box" to a (gasp!) 256KB unit, which could now track gear box data, traction control and anti-lock braking performance. All the data gathered fit on a floppy drive, but it was four times what they had in the past.

Much has changed in the world of F1 since. Given the importance of millisecond improvements in car performance, it is now one of the most data-centric sports (and perhaps even industries) in the world. As such, it is a leader in the big data revolution currently fueling enterprise IT investment today.

Hackland was brought into Williams in 2013, when team performance had fallen dramatically year-over-year. The previous year had been a historic one for the team, winning the Spanish Grand Prix, and optimism was high at the beginning of 2013. The team has a long and successful history in F1, with drivers such as Alain Prost, Nigel Mansell and Damon Hill all winning championships in its cars, and performing well is very important to the team's owners.

But the car failed to perform in 2013, and the owners decided urgent changes were needed. The company had previously seen IT as largely separate from car performance, where costs needed to be minimized. But now it became an important contributor to the cars' performance, and a major digital transformation program kicked off.

Hackland stressed that F1 racing was a constant series of iterations, where car development was always ongoing. But this process was not extended to the factory. Today, 10GB of data is generated every Friday before a race (during race practice), and over a three-day race weekend, the team generates about 60GB of data just on car performance. That doesn't include video analytics, gained from broadcasters and the FIA (Fédération Internationale de l'Automobile) and provided to all teams. And then there's weather data, tire performance metrics, analysis of competitors and their performance and driver comments.

One of the biggest challenges the team faced was that real-time data was not available to the larger team back at their factory and headquarters. This meant the largest brains trust within the company wasn't able to analyze data and suggest changes in the configuration of the car during the race weekend. Video analytics were a huge problem, as they could only be delivered overnight. This meant that findings based on practice could not be implemented in time for the race.

The biggest challenge for Williams -- and indeed any team -- was that many F1 tracks are located well outside urban areas. Connectivity was a huge problem, with latency issues even for conference calls. A key objective for Hackland was to make the team at the factory able to access performance data, run simulations and contribute ideas to the car set-up as if they were track side. Video analytics were a particularly important part of this.

Hackland signed a major networking deal with BT to use it for high-performance networking and cloud services, assuring faster connectivity and data transfer. Williams now has access to a 100Mbit/s MPLS network offering symmetric speeds between Williams' Grove factory and 21 F1 tracks worldwide.

This has not only allowed engineers at the team's headquarters to access and analyze race data in real time, it has also allowed them to be anywhere in the world and still be able to do their jobs. Hackland talked about how Williams' chief engineer was even able to access data while on holiday on a beach in Spain.

It's also dramatically reduced the amount of IT equipment the team has to carry to each race, with much of heavy lifting now done in the cloud and back home, at Williams' headquarters. In fact, the engineers now see the footage faster than live broadcasts on TV. They can now review real-time video analytics on their own cars, competitor cars where FIA regulations permit, and send changes to the track in real time, allowing for improved car performance.

Williams has seen immediate results from these IT investments. The team moved from ninth position in 2013 in the constructors' championship, to third in 2014. More recently, the team has been performing well and is in the top five for the 2016 championship. While the days of Prost, Hill and Mansell may not be around the corner, Williams is doing its best to keep up with the big boys.

The role of video analytics is especially interesting to me, as video becomes more pervasive in the enterprise. This is a good example of productive video being used by organizations to help improve productivity and performance, and highlights the potential for operators to help enable these use cases. (See Defining Productive Video.)

— Aditya Kishore, Practice Leader, Video Transformation, Telco Transformation