When the cars of Sebastian Vettel and Kevin Magnussen briefly kissed on the very first lap of this year’s Spanish Grand Prix, engineers at a special facility hundreds of miles away in Milton Keynes, England, were able to tell whether the defending world champion’s race was over, long before the driver had completed the final three corners of the lap.
It wasn’t: hundreds of sensors in the car sprang to life transmitting vast pools of data, not to the pit lane, but to state-of-the-art processors configured with the latest in real-time data analytics back at the team’s UK headquarters.
By the time Vettel came in for a routine pit stop several laps later, trackside mechanics were already briefed on how to tweak the car’s set-up to prevent further damage and optimise performance and Vettel went on to finish the race in fourth position – one of his best results of the 2014 campaign so far.
In the high-octane world of Grand Prix racing, mere tenths of seconds can separate winners from losers: in the race to provide ultra-fast networks to the eleven teams competing in this year’s F1 championship, the margin for error is one hundred times faster. As providers edge closer to their ultimate dream of delivering data at the theoretical speed of light, mere milliseconds can make a huge difference.
To put that in context, a millisecond is a thousandth of a second: it takes around 80 milliseconds for the human brain to process a thought and about 300 milliseconds to blink.
And following the biggest shake-up in Formula One’s rulebook for more than a decade, every one of those milliseconds count. It typically takes seven years for a car manufacturer to take a new concept from the drawing board to production. Formula One teams must do the job in around five months.
This year, the task is even harder: teams are required to use smaller engines with more complicated gearboxes and they are restricted to using 100kg of fuel or less per race. Scores of other changes to the car’s weight, exhaust, nose, height and front and rear wings have also been introduced.
But the real killer – and the reason why carriers have suddenly emerged as vital lynchpins in the sport this season – is that teams can only test their new cars a maximum of four times over a two-day period before each race. More crucial still, the number of engineers and mechanics allowed to go to each race has been slashed by more than a third. This year, only 60 team officials are allowed trackside.
The upshot of all these changes is that teams have less time and fewer specialists at each circuit to fine-tune their new cars. Which is where providers like AT&T and Tata Communications come in. Matt Cadieux, chief information officer at Infiniti Red Bull Racing explains: “When we show up at a new country – and this season we will have raced at 19 different locations on five continents – it’s really important that we have ultra-fast communications links in place. Each time one of our cars goes out on the track, we are looking to make improvements, so it is vital that we can generate a vast bank of data from the car to the factory back in the UK.”
AT&T in pole position?
In 2011, Infiniti Red Bull asked AT&T for advice on how best to utilise the US carrier’s global network. A year later, the partnership was formalised with a comprehensive multi-year agreement in which AT&T agreed to provide a range of highly advanced network services focussing on an ultra-low-latency bandwidth link.
At the beginning of the 2014 season, the deal was again tweaked to reflect the team’s need for greater bandwidth and speeds. In all, communications links are now around 6.5 times faster than they were before AT&T started working with the team, and the carrier delivered almost half of that improvement during a revamp ahead of the current season.
“The deal is a real game-changer for us,” confides Cadieux. Under the terms of the arrangement, AT&T offers network hosting, cloud services and security services: as well as delivering real-time telemetry, the network carries business traffic and information systems, e-mail and 3D Computer-aided design (CAD) data as well as supporting live video conferencing.
It is primed to handle around 7GB of data in any given one-hour practice session, rising to more than double that during the race itself when, among other things, analysts in the UK will cross-check a driver’s line through a corner with the optimum line in real time and relay suggested improvements where necessary. Over the course of the weekend, the network will handle as much as 100GB of data.
At each circuit, AT&T establishes a Point of Presence (PoP) at the garage in the pit lane, allowing the Red Bull team to plug in and play their own IT systems when they arrive: to the racing team, each circuit effectively appears to have a permanent network infrastructure. The carrier provides on-site support at all hours of the day and night during race weekend and set-up teams leapfrog one another, moving from circuit to circuit, ensuring that the network is fully tested and resilient at each new location before it is required.
“The network AT&T provides is truly mission critical,” says Alan Peasland, head of technical partnerships for Infiniti Red Bull. “Improving the car’s set-up is a never-ending process. AT&T’s network allows us to send information gathered from the car during practice sessions back to our HQ in the UK in real time. Increased network stability and performance ensures that we can send back more data from the car more quickly for analysis. Speed is of the essence.”
At practice sessions ahead of a race weekend, the Red Bull team might typically expect to modify the car’s set-up with ten separate development programmes, each requiring real-time cross-checking with results the team has already run on simulators back at HQ. Once they have an idea whether the data from the car matches that from the development data, engineers can start to optimise the modifications to specific track conditions.
With a vast pool of Big Data on which to draw, very little is left to chance and even less to gut feel: “That’s where the network comes in to its own,” says Peasland. “Our job is to use the data to predict certain outcomes and plan around those.”
Tata enjoys life in the fast lane
Of course, AT&T does not have the F1 market to itself. Last year Tata Communications became the official managed connectivity supplier to the Mercedes Petronas team. Ross Brawn, who retired as team principal at Mercedes in February, described the deal thus: “Innovation is crucial – it’s our lifeblood. Having an ultra-fast and reliable network has become vital to the way we work now – it’s an extremely competitive element of what we do. We are so dependent on the data and support from the factory that if you were to cut it off we would lose our competitive edge. Tata communications gives us the speed to lead.”
The Mercedes Petronas team is based in Brackley, Northamptonshire, close to the famous Silverstone race circuit, the home of the British Grand Prix. Real-time data from more than 150 sensors on the cars, which are relayed back to the factory across more than a hundred channels over an ultra-low latency network, are fed into virtualised versions of the car to highlight potential mechanical failures and other problems. So fast is the link that trackside engineers cannot tell which data is coming off the virtualised cars and which is coming off the actual race cars.
The Mercedes Petronas deal builds on an earlier agreement arranged in 2012, when Tata secured the rights to become the official connectivity provider to Formula One Management (FOM), the company that owns the Formula One franchise. As part of the deal, the financial details of which are not available, Tata became the official web hosting and content network provider for Formula One.
Longer term, the connectivity deal with FOM might have far greater potential to Tata than any efforts to develop an ultra-low latency market among racing teams: it is the vast legion of TV fans that fuels revenues at Formula One, not the number of spectators who turn up at each race. If FOM were ever to ditch its current broadcasting strategy – which relies on delivering race content through a number of free-to-air and pay TV channels – in favour of streaming content direct to the seven million fans who already subscribe to its official website, then Tata’s wholly-owned network infrastructure would assume an even more critical role.
To emphasise the point, Tata recently trialled a new live broadcast service capable of delivering multiple high-quality Jpeg 2000 live video feeds from the Singapore Grand Prix direct to the headquarters of FOM at Biggin Hill, UK. An FOM insider breathlessly likened the demonstration to “the best thing to happen to F1 since the advent of satellites”, but John Morrison, FOM’s CTO, is slightly more circumspect: “We are constantly looking at ways to improve the delivery and number of options for our media partners and Tata’s network will undoubtedly allow us to radically rethink our live event coverage.”
“The possibilities are endless”, echoes Mehul Kapadia, MD of Tata’s Formula One business. “By using our fibre network, broadcasters will have the opportunity to manage content in ways hitherto unthinkable via satellite, giving them options on what video feeds they share with their viewers.”
In the meantime, Kapadia is happy to use the Formula One circus to showcase what he can do for other customers.
“The Formula One championship is a high-profile environment in which to demonstrate the types of services that we can offer other global enterprises.” It boils down to three things, he says: helping customers become more global, helping them become faster and helping them innovate. It typically takes about 30 days to install, test, run and then dismantle the type of MPLS circuit needed to support FOM and Mercedes Petronas at each track.
“We install, test and configure temporary MPLS circuits for them in a week. As a feat of networking, it is equivalent to changing a tyre in under four seconds,” says Kapadia. “If we can do it for F1, we can do it for anyone.”
A point that should not be lost on carriers grappling with ways to make the ultra-low latency market pay.
Setting new records
Until recently, the only customers willing to bankroll ultra-fast connectivity were a very small band of highly secretive outfits on Wall Street known as high-frequency trading firms. These operators had woken up to the fact that they could scalp profits from the tens of thousands of buy and sell orders bouncing around the plethora of new electronic exchanges if they could get to each exchange ahead of everyone else.
The rewards are vast and the risks almost non-existent: in US regulatory filings, one firm – New York-based Virtu Financial – recently confirmed that it had only lost money on one trading day in the last five years. On each and every other day in the 1,238-day period, it had banked more than $1 million.
Like its peers, Virtu is prepared to pay handsomely for the advantages that ultra-low-latency connectivity convey, but the good times may be over for the HFT community amid mounting regulatory concern fuelled by an explosive new book, Flash Boys, by Michael Lewis. In the book, Lewis effectively alleges that HFT firms have used low latency networks to rig America’s stock markets.
Carriers have variously tried to develop low-latency markets in the health and gaming industries in preparation for the day when the HFT cash cow finally dries up, but to little avail. So the evolving need for high-speed connectivity in the motor racing world is indeed a welcome boost.
Crucially, neither AT&T nor Tata Communications will elaborate on the specific latencies that they have been able to achieve between various race tracks and team HQs. Where once carriers trumpeted each millisecond that they were able to shave off specific routes, operators now jealously guard the optimum speeds of their networks, partly because publication sets a benchmark for competitors to attack and partly because future improvements are likely to require the type of significant capital investment that few providers are willing to shoulder, given the limited pool of customers willing to pay hefty premiums for ultra-high-speed connectivity.
Moreover, as with the evolution of the HFT market, there are unique challenges to constantly overcome in the racing community. For one thing, providers must avoid creating a broadband bottleneck between the car’s sensors and trackside receivers as the hunger for data grows unabated. On street circuits surrounded with high buildings, connectivity with cars travelling at speeds of close to 200 miles per hour is tricky to maintain: data bounces off walls, arrives in the wrong order or drops out entirely. But with each passing season, network providers are proving that they have more and more to offer the F1 industry: as an exercise in developing hitherto untapped markets, it is a clear winner.
Crossing the finishing line
If you thought that the relationship between high-speed network providers and the Formula One motor racing fraternity is a one-way affair, think again. As Tata’s Kapadia says, Grand Prix racing is a very strong metaphor for the way enterprise is developing around an industrialised internet: you take information, analyse it and adapt accordingly, significantly enhancing performance along the way. The bigger the bank of information, the better the analytics that come out the other end.
There are very few industries that innovate at the pace of the motor racing community, but the speed at which the sector has grasped the opportunities of Big Data analytics has caught pretty much everyone off guard. This is why IO Data Centres is relying on McLaren Applied Technologies (MAT), the development arm of the iconic racing team, to help it design a new generation of energy efficient data storage facilities.
McLaren has already helped San Francisco’s Bay Area Rapid Transit system design a new Wifi network for passengers, using the experience it has gained from plucking data from speeding racing cars to develop a network that also provides security surveillance. Now it is using its skills in predictive Big Data analytics to help model future demand patterns for data centre services.
Geoff McGrath, MD of MAT, says that the company’s record on the race track helped sway the deal: “McLaren’s ability to analyse Big Data via the use of hi-tech simulators proved a key factor in formulating IO’s decision. IO has challenged us to think differently and to identify ways to improve efficiency, cut energy requirements and reduce emissions.”
“Anyone who has seen a Formula One race knows that McLaren is unrivalled when it comes to integrating outstanding software with the hardware built to receive it,” says Kevin Malik, CTO at IO. “We are extremely excited about this partnership and the potential to make advances in the telemetry of data centre technology.”
