Have you heard about GPS rollover? In recent weeks it’s been widely reported in the news and tech media that a potentially serious event may happen on April 6 involving some GPS systems and devices. This event is due to the fact that GPS receivers not manufactured in compliance with the IS-GPS-200 specifications may see their GPS Week Number count rollover from 1,023 to zero on this date, causing them to interpret April 6, 2019, incorrectly as August 21, 1999, which is 1024 weeks or 19.5 years off time-wise. This will occur because GPS devices utilize an internal counter called the WN parameter that is incremented by a value of one each week. Unfortunately when the GPS system was first designed this counter was defined as having only ten bits, which means it can range only from 0 up to 1023. When the GPS system kicked off with “week zero” way back on Jan. 6, 1980, the counter started incrementing weekly. On Aug. 21, 1999, the counter had reached 1023 and rolled over back to 0, but GPS was not nearly as widely used as it is today so the impact of this first “rollover event” was minimal. Unfortunately with the second possible rollover event now looming on April 6 and GPS devices now widely deployed in industry, government, and personal use, the potential for small- and possibly large-scale disaster looms large.
The good news is that most critical systems that rely on GPS won’t go south on April 6. Unfortunately, there’s also some bad news, and I’ll get to that in a moment. But first let’s look at the scope of the GPS rollover problem, who should be worrying about it, and what you can do about it.
GPS rollover problem: It’s about time
A wide range of systems that are part of the critical infrastructure of many nations rely heavily upon accurate timing information provided to them by the Global Positioning System, a U.S. government-owned satellite-based radio-navigation system operated by the United States Air Force. This includes critical infrastructure for such sectors as financial markets, telecommunications, emergency services, industrial control systems, and power grids.
Many mid- and large-sized enterprises also use NTP time server appliances that utilize GPS timing information to ensure accurate Coordinated Universal Time (UTC) time throughout their network. As those of us who administer infrastructure built using Windows Server are well aware, accurate time is essential for ensuring the reliability of Kerberos authentication in Active Directory. Accurate network time is also invaluable when troubleshooting all kinds of problems involving network communications and resource access as it ensures accurate timestamps for Windows events, transaction log entries, and analyzing network tracing captures.
Some GPS device manufacturers have been releasing software updates for their GPS modules to ensure this problem doesn’t happen — or at least to delay it from happening for another 19.5 years. Many news sources have characterized this upcoming event as being similar to the Y2K bug that was predicted to wreak havoc in computer systems around the world at the start of the millennium on Jan. 1, 2000. That turned out to be mostly hysteria however as governments and industries around the world spent hundreds of billions of dollars preparing for the event by updating old software. But always in search of the next crisis to attract eyeballs, many news sources have been calling out the GPS Week Number count rollover event as the next Y2K crisis, with one security expert attending the RSA 2019 Conference saying he’s “not going to be flying on April 6” because of the potential impact the event could have on airliners, which depend heavily on GPS for navigation (though Garmin indicates there are no known issues with Garmin aviation devices with regard to the rollover problem).
Something to worry about?
How concerned should we be about this problem? If you’re just an ordinary user of a GPS device and you’ve owned your device for about a decade or more then you should check whether the manufacturer has released any firmware updates to address this problem. Newer handheld GPS devices probably won’t be affected by the GPS week number rollover event happening on April 6 — that’s the good news.
If you administer an enterprise network however — and especially if this involves critical infrastructure for government or industry — then you should start by determining which parts if any of your infrastructure may depend upon GPS for reliable operation. Then analyze how your infrastructure could be impacted by GPS timing failure caused by the rollover event. You probably also want to contact your GPS appliance vendor to determine any potential issues with their devices concerning the rollover event. The good news once again is that newer GPS appliances are not likely to go wild on April 6.
There’s also some bad news however. The U.S. Department of Homeland Security’s National Cybersecurity & Communications Integration Center (NCCIC) and National Coordinating Center for Communication (NCC) recently released an unclassified memorandum concerning the April 6 GPS week number rollover event. This memorandum is directed to U.S. owners and operators using GPS to obtain UTC time, and it indicates that while devices that conform to the latest IS-GPS-200 standards should not be adversely affected, tests have revealed that not all GPS manufacturer implementations correctly handle the rollover. More importantly, the memorandum says that some manufacturer implementations interpret the WN parameter as being defined relative to a date other than the original GPS system date of Jan. 5, 1980. Concerning these devices the memorandum states that they “may experience a similar rollover event at a future date.”
If that sounds ominous to you (and it does to me) you may be even more disturbed by what Eric S. Raymond the author of “The Cathedral and the Bazaar” said about this recently in a post he made on the mailing list of the North American Network Operators’ Group (NANOG). Eric, who as a lead for the NTPsec and GPSd projects is an expert in this kind of stuff, indicated in his post that “Everything said in that memo is correct” but added that “it is more normal now than not for GPSes to have a hidden pivot date” meaning that “they will go poof some unpredictable number of weeks later.”
Not may go bad, but will go bad — and at a time the device’s user can’t predict, a time that varies between devices and manufacturers.
How’s that for some bad news?
Featured image: Shutterstock