Re: [Ntp] Antw: Re: Antw: Re: Antw: [EXT] NTPv5 Loop Detection without Stratum

[Miroslav Lichvar <[e-mail]>]

On Wed, Aug 31, 2022 at 11:47:41AM +0200, Ulrich Windl wrote:
> Well, every phone call is a "loop": The microphone records what it hears from
> the other end of the line (and sends it back, where the other end does the
> same), too.
> If the amplification is too high you'll have a never-ending loop of noise; if
> it's low, it will vanish.

It's not just amplification. If you turn on the speakerphone, there
has to be something in HW or SW that is measuring the delay and
suppresing your signal in the loop. That is quite a difficult problem.

If that doesn't work very well, you can hear your echo from the other
end. That is annoying even if it doesn't amplify.

> I think NTP should only forbid the first example of a loop.

So, instead of detecting all loops you want to detect only loops that
lead to an amplifying oscillation? How exactly would you do that and
what would be the advantage?

-- 
Miroslav Lichvar