The specific hardware the IMS uses to contact each decoder - along with virtually every other major racecourse in North America - is called the TranX Pro Multiloop system, manufactured by AMB (no relation to AMD). This system completely accounts for the transponders and the decoders. The software NASCAR uses to manage this system is called TimeGear. It projects the locations of decoders along the racecourse, and monitors the physical condition of each one - not every loop is critical to the operation of the system except the one representing the start/finish line, although the more loops operational, the better. Green boxes on the map indicate loops in nominal working order.
"When the car comes across [a loop]," Worling explained to us, "the transponder in the car gets read by the decoder, and it sends [a pulse] back to the server. As a car comes through, [the software] says, 'Now I know relatively where you're at, and I'm going to put you in order as everybody comes around until you get to the next line, and if you change a little bit, I'll adjust that.' So when we throw a caution, what happens is, everybody gets lined up in their relative positions from the previous loop they were scored at."
NASCAR scoring official James Garr demonstrates TimeGear timing and scoring software, showing the locations of decoders stationed along IMS.
As NASCAR scoring official James Garr demonstrated to us, two screens along the left-side wall are devoted to the cars' running order. When a caution comes out, the bottom display will show the frozen running order, where the cars should be. The top display then continues to show the cars' actual running order, along with notation letting officials know when to notify drivers to move forward or back as they line up behind the pace car. "Then we can see, if they move out a little bit, we can say, 'Hey, 48, you need to back up two spots!'" explained Worling, in a soon-to-be-prophetic reference to Jimmie Johnson.
With the strict enforcement of the frozen field, Garr stated, "[drivers] have no incentive to keep on the accelerator, because by the time the caution comes out, it's already been decided where [they're supposed to go]."
The loop sensors are all embedded at various points in the race track. As the track ages - as asphalt tends to do very quickly - the track surface itself can actually weep. This might move the sensors an inch or two over the years. Doesn't that affect the scoring? "No, because it's all relative," said Garr. "All we're trying to do is figure out which car is in front of the other car. So I don't really care how many inches this car is in front of that car. When a caution comes out, I'm not concerned about the start/finish line. I'm concerned about the last line that they passed."
The complete network bandwidth, Worling stated, has the usual Ethernet bandwidth of 100 Mbps. That's more than enough, he said, because the data each decoder sends through the loop - complete with car identifier and timestamp - is only about 60 K per packet. At full streaming, he said, bandwidth consumption only runs about 2 Mbps, and that's counting the separate feed to officials who continue to occupy their traditional locations inside the pagoda.
At the end of the race, the entire database collected from every point in time is transmitted to NASCAR's headquarters in Daytona Beach, Florida. It's a proprietary, flat-form database that can consume about 200 MB of hard drive space - as much as 1 GB for a race with more decoder data or a longer road course - but when sent in compressed form consumes only about 15 MB. The connection used to send this file is not the Internet - rather, it's a corporate VPN, the tunneling data for which is never shared with the outside world. Having the race data hacked is something NASCAR never wants to experience.