justinkendra - Fotolia

Microservices: An integral part of tracking Super Bowl player data

While they may seem worlds apart, microservices and the NFL have come closer together than ever. In fact, they are a key part of what we know about the players in Super Bowl LI.

Of all the bets being made over Super Bowl LI, there probably aren't that many of those betting that most NFL players know what a microservice is. But while that may be the case, in the last few years microservices have become an integral part of player-tracking technology that gives NFL fans, officials, coaches, staff and broadcasters access to more statistics about the game than ever before -- and at an unprecedented rate of speed.

Zebra Technologies, a provider of technologies that create real-time visibility into business operations, has been working with the NFL since 2013 to fit stadiums, players and gear with tracking technology that gives those both inside and outside the league never-before-seen insights into player behavior on the field. And according to Jill Stelfox, vice president and general manager of location solutions at Zebra Technologies, microservices play a critical role in acquiring, processing and distributing this data quickly, and will continue to do so as they track this year's Super Bowl player data.

How it works

Jill Stelfox of Zebra Technologies says that microservices are a huge part of what makes their player-tracking technology work.Jill Stelfox

Since 2013, Zebra Technologies has used its collection of data tracking hardware, Dart Ultra Wideband, to track players. This system relies on RFID sensors placed around the stadium and in every player's shoulder pads -- it's part of the official NFL uniform. These include 22 receivers in the stadium, two RFID tags on every player, two tags on each referee and a tag in the ball. The tags, Stelfox said, track up to 25 times a second and deliver information, including latitudinal and longitudinal data, in about half a second.

With that data, Stelfox and her team are able to use their sports tracking software, MotionWorks, to calculate things like speed, distance, closing distance, routes and formations. They then quickly add what they call "eventing data" -- data that helps translate the data into a football context -- and transmit this info at a head-spinning rate to NFL officials, coaches and fans.

"That information goes out in under two seconds to a number of users including the NFL, but it also shows up on Xbox One and television," Stelfox said. "So there are a lot of places that the data goes after that."

RFID sensors are now placed in every NFL player's uniform to track a variety of metrics.
RFID sensors are now placed in every NFL player's uniform to track a variety of metrics.

How microservices make it happen

According to Stelfox, microservices play a big role in gathering and distributing game data in a timely manner, and will continue to be one as they track Super Bowl player data this year. What microservices allow for them, Stelfox said, is speed. Microservices allow them to capture information about the play as it is unfolding, information they can send to broadcasters immediately upon the end of a play for use in TV replays.

The reason microservices make this happen, Stelfox said, is because of the distributed compute power they enable. The ability to distribute where logic is run instead of running all logic together is what makes it possible to track so many metrics simultaneously and at such a fast speed.

What microservices allow for us is speed.
Jill Stelfoxvice president and general manager of location solutions, Zebra Technologies

"You've got twenty-two players, you've got all kinds of plays happening," Stelfox said. "You wouldn't be able to deliver the information as fast as you need to if you don't spend a ton of time looking at how you architect this thing, [including] where you put compute."

But the use of microservices is nothing new for Zebra, Stelfox said. According to her, Zebra had already been putting microservices to use to deal with fast-paced safety situations in warehouses. Luckily, she said, there shouldn't be a Super Bowl player who moves faster than a forklift.

"We use the same logic in terms of distributing compute power to do that kind of thing where we can stop a piece of equipment from doing something that is harmful," Stelfox said. "So that logic we had already, it's just applied in a different way."

But while making this technology work for the NFL is mostly a case of reapplying their existing logic, Stelfox said that there was a notable challenge: translating the notion of warehouse business rules to apply to the NFL rulebook. This involved learning about even the most fringe cases of football situations -- something that is essential to delivering information quickly.

"There's the odd play that might occur and you're like 'Oh, what's that?'" Stelfox said. "But you can't ask 'What's that?' if you're streaming in half a second ... the computer has to know what it is."

Simple insights from data change the game for NFL players

Player tacking technology implemented by Zebra Technologies is able to provide fans with a new dose of football-related stats, such as how fast a player is running or what their vertical leap was during a play. But while these might seem like arbitrary stats from a game perspective, the data that Zebra provides players and coaching staffs has a big impact on the game, including pinpointing the attributes of a strong player and forming training strategies in a timely manner.

Jill Stelfox, vice president and general manager of location solutions at Zebra, said there are two statistics their technology has been able to gather that teams find particularly interesting and desirable to know. The first is closing distance -- as in, how fast a defensive player can close in on an offensive player. The second, she said, is yards of separation. This refers to the average yards of distance a wide receiver is able to put between themselves and a defensive player by the time a reception is made. The average player, she said, can achieve a yard and a half difference, while elite players have been seen to average 3.9 yards of difference -- shedding more light on just what makes these players so successful.

The physical training staff of NFL teams, Stelfox said, is also gleaning invaluable results from the use of this data tracking technology when it comes to helping players recover after games and prepare for the next one. For instance, using the data from Zebra's player tracking, staff members are able to see exactly how many reps a player ran during a game, something that traditionally wasn't always obvious to staff until a review of game tape. Because trainers are able to see that info a few hours after the game rather than the next day after tape review, they can have that much more time to put together an individual player's training plan based on their workload the previous game.

"When the player comes into the locker room on Monday, you can already have devised a plan for recovery," Stelfox said. "So it lets you just get after it faster."

Another piece of data that Stelfox finds particularly fascinating -- and useful for trainers -- is the total number of yards run during the game. Players do a lot of unrecorded running in games outside of recorded plays -- for instance, running out to start a drive, running onto the field for substitutions and running in a play that is reversed. And though those distances are not recorded officially, knowing exactly how much a player ran during a game can be a key part of putting together recovery and training regimens.

"In football we usually only record the total yards during the actual play itself as a statistic," Stelfox explained. "But there is a tremendous amount of running that's not counted ... it ends up being almost 3-to-1 in some cases."

An unexpected hardware challenge

The technical challenges of working with the hardware and software that goes into the player tracking systems were minimal, Stelfox said. According to her, the biggest challenge arose from the fact that the stadiums are not always owned by the NFL or its respective teams. Often they are owned and operated by local stadium authorities or a private party, which turned out to be an issue when it came to scheduling the installation of hardware. Usually the companies they work for set aside time for the project, but in this case they had to work around other stadium events.

One unexpected individual turned out to be a consistent roadblock when they were preparing for the 2014 season, Stelfox said: Taylor Swift, who happened to be on tour at the time. Everywhere they went, they would have to leave for up to eight hours while the pop sensation practiced for an upcoming show.

"It seemed like we were following her concert," she said. "We'd go to Atlanta, and she'd be in Atlanta; and then this week, New York, and she'd be in New York."

Microservices allow Jill Stelfox and her team to transmit game data to those who need it in seconds.
Microservices allow Jill Stelfox and her team to transmit game data to those who need it in seconds.

Passion and speed

Although the program was designed to help players, coaches, NFL officials, broadcasters and fans, the developers at Zebra achieved their own benefit from this project as well. According to Stelfox, the developers on her team were able to take something they loved already, football, and apply that passion in a meaningful way to their software development work.

"We have software developers that love and appreciate football," Stelfox said. "And I think this really does allow software developers who have passions for sports to marry those two things together in a really fun and interesting way."

Stelfox also remarked that the demands for fast information are really representational of the software development scene as a whole, an environment that continues to make speed a priority.

"The way software development works is faster, faster, faster, faster," she said. "And I think being able to take half a second down to 100 milliseconds, speaking through how you architect these kinds of robust solutions, is just ever more complicated."

Next Steps

How to watch the Super Bowl on your tablet

How the NFL uses Wi-Fi analytics to keep Super Bowl goers happy

SAP HANA steps up to quarterback Super Bowl apps

Why a Super Bowl survey shows that app performance matters

Dig Deeper on Distributed application architecture