Watch Out for the Leap Second – Again!
This is another in our irregular series of guest blogs, this time from Mitch Narins at Strategic Synergies, LLC.
The old adage tells us that “Time is money,” but in today’s world, time is so much more. Precise time is the basis for safety, for security, and for so many of the benefits that our modern infrastructure systems provide. Without precise time, many of the services upon which we rely would be unavailable – or even worse, potentially lacking integrity and thereby posing unforeseen dangers.
Precise time is critical. Errors of billionths of seconds (nanoseconds) equate to position errors in feet.; errors of millionths of seconds (microseconds) will impact telecommunications and power grids; and errors of thousandths of seconds (milliseconds) impact financial transactions.
So if these extremely small fractions of a second are so important, what would happen if suddenly, even in a planned manner, time were to “jump” a full second? What if we decided as a society that we should have a 61-second minute instead of the normal 60-second one? What if the timing receivers upon which we rely didn’t properly handle the change?
This year, once again, we get to see (pardon the pun), “in real time,” the effects of adding a leap second at the end of calendar year 2016.
Leap seconds have been added periodically over the years to compensate for the slowing down of the Earth’s rotation to keep precise scientific time in synch with observed astrological time. Groups continue to debate the wisdom of doing this, but that is a subject for a different discussion. The fact remains, this year will be one second longer than normal years – so why is this a problem?
On July 19, 2016, the Civil GPS Service Interface Committee (CGSIC) announced the following:
Yesterday, 18 July, the Air Force entered, into the GPS MCS, a Future Leap Second to become effective on 31 December 2016. As Navigation Uploads are/were performed over the following ~24 hours, GPS satellites will, one at a time, begin broadcasting this Future Leap Second, along with its date of effectivity and the Current Leap Second count, in accordance with IS-GPS-200. When specific satellites will begin broadcasting the new information will depend on the contact schedule over the next ~24 hours, which is dynamic and subject to real-time operations. When specific users will begin seeing the new information will depend on a number of factors, including satellite visibility and user equipment design.
In short, this means that (1) the GPS constellation is now broadcasting information to GPS receivers to “prime” them for adding the leap second at the end of calendar year 2016, and (2) the last minute of the year, which starts on December 31, 2016 at 23:59:00Z will now end at 23:59:60Z (instead of 23:59:59Z), meaning that the start of 2017 on January 1, 2017 at 00:00:00Z will be delayed by one full second.
Again, if this is planned so far in advance, why is this a problem? Good question! Most GPS receivers will sail through the change without a problem. But this hasn’t been true for all receivers in past leap second years. Some were improperly programmed and subtracted a second instead of adding one. Other receivers were not designed to accept the Future Leap Second notice and failed to implement the required correction. Still others weren’t able to deal with the Future Leap Second message (Note: USCG Navigation Center reports that this has already happened, as a result of the 18 July Air Force notice!).
In the past, some organizations have briefly taken their systems down during the leap second transition to preclude problems, thus protecting against integrity problems, but accepting the “cost” of reduced capacity and/or efficiency. “Why should a loss of service at midnight have such an impact?” you might ask. The answer – because the leap second insertion is at midnight Zulu (Greenwich) time – that’s 19:00 (7:00 pm) on the East Coast of the US; 16:00 (4:00 pm) on the West Coast. Lots of stuff happening during rush hour and the afternoon push!
So if you’re a user of precise time supporting safety, security, or economically critical operations, what should you do now and what should you do in the future? First, you need to determine the source of your precise time (most probably a GPS receiver). Then, whether it is properly designed and programmed to correctly interpret the Future Leap Second notice — and to insert it correctly. This could require contacting the manufacturer(s) of your receiver(s). You might find you need a software/firmware upgrade, or you might even need to have your receiver(s) replaced.
You can also ensure that you have multiple sources of precise time, so that a failure of one could be mitigated by continued service from different, independent, and resilient sources. Not only is this a good idea for ensuring survival of a leap second insertion, but it also ensures that delivery of precise time is not a single point of failure for your enterprise.
On a final and different note, a few months ago, some receivers suffered a GPS 1024 Week rollover problem, causing them to stop operating (i.e., they incorrectly “reached” the 1024 GPS week many years before the actual 2019 event. NIST has now found that Motorola Oncore UT+ receivers with older firmware (versions 2.2 and earlier) will suffer this same GPS week rollover problem on August 14, 2016. Hopefully users of these receivers will upgrade their firmware or replace them before they suffer any loss of service. Just another reason why seeking resilient alternative solutions is a prudent strategy for all precise time applications.
 GPS provides a correction message that translates the GPS internal navigation time scale into Coordinated Universal Time (UTC), which is the accepted time standard in most major industrial nations. GPS time is limited by a 1024-bit week number representation limit in the GPS navigation message, which started on Jan 6, 1980. The GPS epoch repeats every 19.7 years (1024 weeks) and it is up to the user receiver to solve this ambiguity.