[Ntp] Antw: Re: Antw: Re: Antw: Re: Calls for Adoption -- NTP Extension Field drafts -- Four separate drafts

["Ulrich Windl" <[e-mail]>]

>>> Harlan Stenn <[e-mail]> schrieb am 12.09.2019 um 13:03 in
Nachricht
<[e-mail]>:

> 
> On 9/12/2019 3:15 AM, Ulrich Windl wrote:
>>>>> Miroslav Lichvar <[e-mail]> schrieb am 12.09.2019 um 11:46
in
>> Nachricht <20190912094603.GN21704@localhost>:
>>> On Mon, Sep 02, 2019 at 05:52:33AM ‑0700, Harlan Stenn wrote:
>>>> The thing is that NTP is a complete specification with a goal of time
>>>> synchronization that includes *known behavior in the face of impulse
>>>> response*.
>>>>
>>>> If one uses different algorithms with different impulse response
>>>> behaviors, this can trigger oscillations and even unstable tracking (the
>>>> wheels fly off the bus).
>>>
>>> This was discussed a bit on the last meeting.
>>>
>>> Does anyone have an example of a PLL that is unstable when cascaded
>>> with a different loop? If something works well enough to be stable
>> 
>> I think it was the old Dave Mills consensus like: "It's only NTP if it
>> provides the same output for the same input". So if you think of some
>> improvements, they still would not be allowed (I had some discussions
while
>> implementing PPS in the Linux kernel)
> 
> It's not only about instability.
> 
> It's about known and agreed-upon impulse behavior.

But is the overshoot really well-defined NTP behavior? Are you saying that
anything that doesn't overshoot is not NTP?

> 
> -- 
> Harlan Stenn, Network Time Foundation
> http://nwtime.org - be a Member!
> 
> _______________________________________________
> ntp mailing list
> [e-mail] 
> https://www.ietf.org/mailman/listinfo/ntp