Re: [Ntp] CLOCK_TAI (was NTPv5: big picture)

[Philip Prindeville <[e-mail]>]

> On Jan 6, 2021, at 3:36 AM, Martin Burnicki <[e-mail]> wrote:
> 
> Philip Prindeville wrote:
>>> On Jan 5, 2021, at 5:45 AM, Hal Murray <[e-mail]> wrote:
>>> 
>>> A copy of the leap seconds file is included by the Time Zone project and 
>>> distributed by IANA.  Most distros are already setup to keep the time zone 
>>> info up to date.  It should be simple for the leap file to go along for the 
>>> ride.  Debian already does it.
>> 
>> That seems like a good argument to NOT include it in the protocol, if it’s already available out-of-band.
> 
> If you force using TAI for NTP but don't include the relevant
> information in the NTP protocol, this means that you are generally
> unable to operate NTP properly unless you always take care and maintain
> the additional information.


That’s a non sequitur.  It can be a requirement of NTPv5 that the host environment includes a leaping database, daylight time calendars, etc. to put canonical time into both UTC and local time...


> In the past there was no such requirement, and I think it's a drawback
> if you add such a requirement now.


Like I said, it’s an effort to remove inherited encumbrances from NTP that never should have been there:  if we add any complexity at all, it should be to its core requirements (like security).

-Philip


> Martin