[LEAPSECS] A standard for leap second smearing

[Stephen Colebourne]

On 28 September 2016 at 12:35, Martin Burnicki
<martin.burnicki at burnicki.net> wrote:
> I think in general we need to distinguish if
>
> - smearing is done by a server, so that it hides the leap second from
> all its clients, and the clients don't even become aware of the leap second
>
> - smearing is done by a client, which receives a leap second warning and
> thus is aware of an upcoming leap second, and can either slew its own
> system time over the leap second (that's what ntpd for Windows does) or
> let the OS kernel handle the leap second (stepping the system time back
> by 1s by default) and just hide this to the user level.

This seems key. Poul-Henning Kamp suggests that ntp clients would deal
with smearing badly, taking over an hour (4096s) before making any
changes. Whereas, if the OS kernel receives the correct ntp signal but
publishes the smeared version to users, its just local maths (but
depend on how far in advance the leap second is advertised).

Maybe there are two variants here? - one for the ntp server/client
boundary (needs a slow transition) and one for the OS kernel/user
boundary (needs a fast transition).

Stephen