You might never have heard of Galileo, the European Union's GNSS, or Global Navigation Satellite System. That's because your phone probably uses the US system known as GPS or Global Positioning System. There are actually four GNSSs: GPS (US), Galileo (Europe), GLONASS (Russia), and BeiDou (China, and not to be confused with Baidu the search engine). GPS and GLONASS are fully operational. Galileo and BeiDou are planned to be fully operational in 2020. In the case of Galileo, that means 30 satellites, consisting of 24 operational ones and 6 spares. There are currently 22 operational. The European system went live in 2016, and is owned and operated by the European GNSS Agency headquartered in Prague.

Outage

Well, European nations not having to rely on GPS didn't exactly go according to plan recently. Galileo went down on July 11, and stayed down for a week. I was actually in Europe for my annual hiking holiday with a few friends (17th year), but I didn't notice since my smartphone uses GPS. Even devices that use Galileo will fall back to GPS when Galileo fails, so it wouldn't be obvious in any case. Luckily, Galileo failed totally and so all receivers knew it was down. It is not hard to imagine failure modes where it simply provides bad data and so appears to be working, just locating everyone in the wrong place. According to Galileo's website, there are over 100M devices that receive (or can receive) the Galileo satellite signals.

It is unclear exactly what caused the outage. It seems that it was a problem in the ground stations, which are at Oberpfaffenhofen near Munich in Germany and Fucino, 80 miles east of Rome in Italy. The only official description so far is:

The technical incident originated by an equipment malfunction in the Galileo control centers that calculate time and orbit predictions, and which are used to compute the navigation message. The malfunction affected different elements on both centers.

Wired's report has a little more detail from a Turin professor:

"From the operational satellites the signals were good, the time was good, but in the messages the satellites send the positions were not updated," says Fabio Dovis, a researcher in the satellite navigation lab at Politecnico di Torino University in Italy. "We tried replacing that part of the message with information on position taken from other reliable sources and you could get the correct position. So it was working, it just seems like there was a problem in the control system that was not able to update this information."

Vulnerability

The thing this outage brought home to me is just how much we rely on GPS these days. And when I say GPS, I mean specifically GPS. There may be 100M devices using Galileo, but most of them (maybe all) can fail over to the US GPS system if necessary. But there are about 3B smartphones, many (most?) of which use only GPS. There are other dedicated GPS devices in cars and other places, too, but these are dwarfed by the number of smartphones.

I notice that young people (under 30) seem to only be able to drive using GPS and a map program such as Google, Apple, or Waze. Since they've always had map programs since they learned to drive, they never really get familiar with the route to anywhere, even places that they have visited many times, and as a result can't cope without their phone as a crutch. Of course, all of us use it for finding unfamiliar addresses or finding our way around cities we're not familiar with.

All sorts of services such as cellular networks rely on GPS, too. Not for location, of course, the networks know where their towers are, but for very accurate synchronized timing. The timing is so precise that GPS satellites have to correct for relativistic effects, despite the satellites traveling really slowly compared to the speed of light.

They automatically handle things like leap seconds, too, which are required to correct differences between UTC, the time we use, and solar time, which should have the earth at exactly the same position around the sun every year. The last one was inserted on December 31, 2016 at 23:59:60. Yes, that 60 is correct since with the leap second there were 61 seconds in that minute. There have been various proposals to get rid of leap seconds and just have the time drift slightly, but so far these have been either rejected or postponed (the next decision point is 2023). The organization that controls leap seconds is IERS, the wonderfully named International Earth Rotation and Reference Systems Service (yes, the abbreviation doesn't really match the name). The next potential leap second would be on December 31 of this year, but just a couple of weeks ago it was announced by IERS that there would be no leap second inserted...so New Year's Day won't start a second late in 2020.

More Information

There are articles on the outage in The New York Times, Wired, Aviation Week, and presumably elsewhere. 

You can track the current status of the service on the Galileo website.

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.