NTP group moves to drop leap seconds from NTPv5 draft

Miroslav Lichvar has proposed stripping leap-second support from the Network Time Protocol version 5 draft, arguing that a new BIPM resolution makes the mechanism obsolete for hundreds of years.

The International Bureau of Weights and Measures raised the allowed difference between UTC and UT1 from under one second to one hour. That change removes the near-term need for leap seconds, including the previously feared negative leap second that testing had shown many systems handle badly. Positive leaps already forced awkward workarounds such as leap smearing; under the new rule UTC stays monotonic and linear for the foreseeable future.

Lichvar’s draft changes would delete the leap-seconds section, force the leap indicator to zero, and drop the leap-smeared timescale. UT1 would matter more for applications that still need Earth rotation. If leaping ever returns while NTPv5 is in use, it could be added later by an RFC update.

Responses on the NTP list were supportive. Martin Langer of Germany’s PTB called leap seconds obsolete. David Venhoek backed removal and suggested marking the leap-indicator field reserved. Ruben Nijveld noted that any future leap scheme would likely need a different design anyway. Doug Arnold of Meinberg explained the background: Earth’s temporary speedup had raised the risk of a never-before-seen negative leap, and the one-hour tolerance simply defers the problem for centuries.

The BIPM decision text is recorded as Resolution 5 of the 2026 General Conference on Weights and Measures.