[Ntp] Antw: [EXT] Re: NTPv5 draft

>>> Steven Sommars <stevesommarsntp@gmail.com> schrieb am 30.11.2020 um 21:32 in
Nachricht
<CAD4huA79u3NkR8LS96Gqgs58mJguoN7=p=CJnitzgh_RzNWCVg@mail.gmail.com>:
> Agree, these fields could use better descriptions or (better) replacement.
> 
> One manufacturer (LeoNTP) lowers precision to indicate reduced accuracy and
> claims to obey the NTP specification.

Actually if you consider precision to be a "runtime value" (as opposed to a "design value") it's obvious that it is actually variable. However the  overhead to update the value was considered too heavy in NTP design, I guess.

> 
> Should a stratum 1 NTP server report root distance=0  ?   National labs
> differ.

Would that mean that delay and dispersion are constant?

> 
> Using NTP short format the smallest non-zero root dispersion is 15
> microseconds.  One can argue for finer granularity

Not just there: stability (clk_wander) could also need more digits. clk_jitter is another candidate (it looks like a constant here for GPS). Maybe more. Some are heritage of microsecond resolution of NTPv3.

Regards,
Ulrich

> 
> Steve Sommars
> 
> 
> 
> 
> 
> On Mon, Nov 30, 2020 at 2:03 PM Kurt Roeckx <kurt@roeckx.be> wrote:
> 
>> On Mon, Nov 30, 2020 at 08:12:06PM +0100, Dieter Sibold wrote:
>> >
>> > 3. Traceability
>> >
>> > It would make sense that the v5-packets optionally provide information
>> about
>> > the uncertainty of the timestamps taken. These formally for establishing
>> > traceability.
>>
>> The NTPv4 packet contains 3 fields that's related to the
>> uncertainty: precision, root delay, root dispersion. I'm not sure if
>> anybody is using the precision for anything, but it also has an
>> effect on the uncertainty.
>>
>> I'm wondering if the combination of those 3 fields is useful, and
>> if we need other fields to replace them. We at least need some
>> specification on how to calculate the values. Is the root distance
>> a mean, filtered or last value? Does the dispersion field have a 95%
>> (k=2) confidence interval?
>>
>> > Additionally, in order to maintain traceability during the
>> > time period in which leap smearing is applied the client needs to obtain
>> the
>> > necessary information to calculate the offset between UTC and smeared
>> time.
>> > This also is mandatory to maintain traceability.
>>
>> I think the document, at least the version in git, has an
>> extension for that.
>>
>>
>> Kurt
>>
>> _______________________________________________
>> ntp mailing list
>> ntp@ietf.org 
>> https://www.ietf.org/mailman/listinfo/ntp 
>>