Civil time is based upon coordinated universal time (UTC), which is synchronized to the length of the day. However, the earth’s rotational speed is slowing over time, i.e., the length of the day is increasing. In order for UTC to keep pace with the growing length of the day, UTC is delayed at pre-announced dates by one second. This delay is effected by adding a “leap second” at the end of the day of each pre-announced date. Although leap second correction achieves synchronization between UTC and length-of-day, computer systems that presume a smooth, always-forward movement of time may malfunction when such corrections are carried out.
“Leap smear” techniques are used to achieve synchronization between UTC and length-of-day while keeping smooth the flow of time for computers. Rather than add a second at a particular time of a day (e.g., at the end of the day), each second of the date (or another period) over which leap-second correction is carried out is dilated. For example, for the 86,401 atomic-clock seconds on the day of leap-second addition (with the leap second being excess over a normal day of 86,400 seconds), each of the 86,400 seconds of the day is adjusted to be of duration 1.00001157407 (86401÷86,400) atomic-clock seconds. Leap-smearing works well for computer applications other than those that require high sub-second accuracy. This disclosure describes techniques that allow use of leap smearing to applications that require sub-second accuracy.
Shields, Michael, "Leap-smeared representation of time for high-accuracy applications", Technical Disclosure Commons, (December 07, 2016)