[ntp:questions] Leap second functional question

Martin Burnicki martin.burnicki at meinberg.de
Tue Feb 19 11:41:48 UTC 2008


David Woolley wrote:
> But you can safely have it set for most of the previous six months and
> the following six months, whereas the questioner is assuming it must be
> cleared immediately the leap second has been implemented and not set
> more than a very short time in advance.  (It certainly has to be set
> hours in advance, because some clients may not have polled within an
> hour, and each stratum can extend the propagation delay of the setting
> of the flags.)

Currently the preferred dates to insert a leap second are end of June or end
of December. Usually the IERS publishes a new "Bulletin C" a few days after
one of those dates, indicating whether a leap second would be introduced at
the next of those dates.

So if a leap second shall be inserted then it is indeed announced by the
IERS about 6 months before the leapsecond will occur. However, the bulletin
explicitely mentions the exact date at which the leap second will be
inserted.

E.g. for the last leap seconds which have occurred the GPS satellites have
started to announce the leap second shortly after the announcement had been
published by the IERS. However, the data sets of the GPS satellites include
a GPS week number plus the day number in that week, at the end of which the
leap second is to be inserted. So even if a leap second for December is
announce in July, it's clear the leap second must not be inserted at the
end of September.

Having an exact date of the leap second available (which is also true for
the NIST leap second file used by ntpd) is different than just having an
announcement flag without a specific date. 

In my opinion a protocol which is unable to specify a certain leap second
date should not start to send a leap second announcement flag before the
month at the end of which the leap second shall occur. This applies also to
the leap second indicator (the flag) in the NTP packet.

Relying on specific dates like end of June and end of December (as done by
ntpd), or even additionally end of March or end of Septermber is a bad
practice. The only plausibility check should be for the end of a month.


Martin
-- 
Martin Burnicki

Meinberg Funkuhren
Bad Pyrmont
Germany




More information about the questions mailing list