Tour de France 2019: How Real-Time Analysis Saved the Race
By now you will have all seen the news articles and YouTube videos. This will go down in history as the only stage ever to be cancelled part way through… I was there, saw the landslide up close, and also witnessed NTT’s real-time data literally saving the Tour de France. (This is not widely known, but I’m telling you it happened!)
Let me give you a bit of the background to this once-in-a-lifetime project.
Tour de France behind-the-scenes
A couple of days ago, I visited the South Africa-based NTT (formerly Dimension Data) Team at the penultimate stage of the race. NTT not only own a team in the Tour de France, but they are also a technology partner to the event—and a technology partner with Riverbed. Our Alliance Manager was kind enough to give up a single ticket for me to attend the event. Over the last 3 weeks they have invited and hosted 100s of their customers, giving them privileged behind-the-scenes access. I found myself in the company of some of the most influential people in South Africa.
It was an early start at 0500 on Thursday. My plane was second out.
Tech stuff to whet your appetite
Riverbed has been partnering with NTT/Dimension Data to monitor an awesome new platform called “Race Center.” Race Center provides every detail possible on the status of each stage including where, who, why, and what might potentially happen in the race. The data behind the platform is so high definition, that it can almost predict which rider and team will win each stage. What it cannot control is telling us who might crash, be pushed off, or how the weather might impact each stage (like it did last week!). Everything from sleep patterns, diet, fluid and energy levels, performance from similar stages from each rider over the last 6 years is all taken into account.
Dates: Tour de France ran from 6-28 July 2019.
Platform: Based on a Windows Server Farm running API and Web tiers running .NET Sitecore.
Volume: Monitored 859 million transactions during the event.
Usage: Global audience from 230 countries. The USA was the single largest audience with 62 million transactions throughout the event.
Storage: All of the above monitoring data was stored in only 3.5TB of disk—every single transaction with full troubleshooting details.
Portal: Riverbed provided NTT with 5 or 6 monitoring dashboards that were displayed in the NTT Tech Truck. These dashboards were delivered via the Riverbed Portal running as SaaS in Azure.
What did I get up to?
On Thursday we were invited into the tech bus and exhibition center. This was a fully kitted out mobile demo center showcasing Race Center and some of the new cutting edge tools that are being developed.
Part of the demo was showing us how they track each rider using GPS and with 1 second granularity. The helicopter in the race was collecting all the riders’ data and forwarding it using satellite communication to the data center. The session lasted over 2 hours and was the first chance to video call the race tech bus and speak to the data analysts who were monitoring the race, controlling the twitter/social feeds, and analyzing performance with Riverbed AppInternals.
Friday morning we all had a 0600 alarm call. We were invited to ride the real race bikes from 2018’s event. Each bike was made of carbon and weighed as much as a can of coke. The cost of such bikes range from £15K to £18K. I was trying to sit this one out as soon I found out that tight lycra was compulsory and the ride would last around 21 kilometers. (I have never been on a race bike, never worn lycra and certainly never peddled 21k before!)
So glad I did it!
Le Tour fever
Next we were packing up and heading to the start of Friday’s stage. This is where I met all the team behind the scenes. Fitness experts, dietitians, logistical experts, mechanics, drivers and, of course, the team’s owner, Doug Ryder. This is where we learned about the day-to-day routines that each rider follows to remain at peak fitness.
FACT: Some riders only consume water and rice during the race (with energy pills/bars). Some riders are permitted a glass of red wine too.
The race was about to start, and I had filled up my bag with as much free merchandise as I could get my hands on. It was time to head to the skies and watch the race commence from a helicopter—never been in one before and was not a good time to mention my fear of heights! This was a 20-minute flight that would take us 10 miles in front of the riders. Then we would follow the race route in a car.
We traveled the remainder of the route in 4 AMG Mercedes driven by rally car drivers, passing through many towns and villages where Le Tour fever was in full swing. At every inch along the route, cars, people, camper vans, and flags were everywhere. I’d never experienced anything like it and people waved at us like we were royalty. At points we stopped due to the people congestion—the roads were rammed with spectators trying to find a spot to watch the riders! We shared high fives with the kids and families who thought Christmas had come early.
Caught in the storm
As we traveled up the Alps getting higher and higher, the temperature was dropping very quickly. At the start, the temperature was 36 degrees.
By the time we got to the highest point of the route, it had dropped to 12 degrees and a storm appeared from nowhere. Hail stones the size of sugar lumps hammered the cars and we literally had to stop due to lack of visibility and for our own safety. It was at this point a landslide blocked the road behind us, right in between car 2 and 3. (I was in car 2.) If we were 3 seconds slower, our car would have been sliding down the side of the mountain!
Our drivers phoned in the news to the HQ of the tour. This landslide was the cause of the cancellation and the first time ever the race had been stopped. News slowly spread and thousands of spectators up the mountain were totally deflated, the mood changed, no more waving or cheering, not even a smile. No race, no riders, no excitement and, to make things worse, the temperature was plummeting, and they were all dripping wet. We reached the finish of the race, and headed to the NTT data bus where we had planned to see out the final stages of the race in real time. As we walked into the bus there was some exciting news developing. All of us on the bus, guests and execs included, were told to leave immediately. We were confused, wondering what in the world was going on. Two of our cars were still stuck on the mountain, the race was cancelled—but what else?
How a high definition platform for real-time analysis saved the race
The race officials that cancelled the race had decided to take the important rider times from the highest point of the race as all riders had completed the climb and were heading downhill.
However, the time sponsor was unable to provide the times for each rider because it was the only point along the whole 130km route that did not have the checkpoint for time. Race officials were in a state of panic. Without the times, they could not firmly indicate who was in the lead. They would have to go back to the previous marker and this would cause chaos among the riders and teams. The officials turned to NTT’s Race Center data for help. Using exact longitude/latitude data and time stamps, NTT was accurately able to provide times at any point along the route to a single second. NTT’s data had just turned into GOLD! It quite literally saved the day!
When we eventually reentered the bus, our two remaining cars had finally caught up. We chatted about the importance of 1 second data, and how Riverbed AppInternals collects real-time Race Center transactions at the same granularity.
In 2015, there was a huge crash in the race. It was also the first time GPS sensors were being used with 1 second data. The graphs had shown sharp increases and declines as the riders flew all over the place. In 3-4 seconds, over 20 riders had crashed. Without 1 second data, they would not have been able to tell who was out and who was in the race. 5 second granularity would have shown 20 riders had stopped but would not have shown why. With 1 second data, they now know with fine accuracy when a crash happens based on the sharp increases and decreases. Not all the riders are tracked by TV cameras so, since 2015, officials use the same patterns in the data to track the safety of riders too. Familiar story, right?
With our two cars now caught up, the riders then started to arrive in their cars. It was a strange end to an amazing few days. We ended the day a little deflated, but did manage to find a lovely Irish Pub to spend some time while we waited for the roads to unblock.
See our technology in action
For anyone following pro cycling, NTT’s web-based Race Center application is revolutionary. Combining live data, video, images, social media feeds and commentary, it gives fans, teams and organizers an unparalleled view of the action. If you missed it, do check out last month’s webinar with Mark Burton, from Riverbed, and Tim Wade, from Dimension Data/NTT, to find out how they’re using a high-performance platform to deliver superior experiences with quick access to real-time data with 100% accuracy and no blind spots.