Harlan Stenn wrote:
> Richard,
> 
> It would be *better* if his servers showed less delay, but I don't see a
> problem there.
> 
> 
> And given he says that machine is the local master for his network, the
> choice of a local refclock is the traditional way for his network to stay in
> time in case he loses internet access.
> 
> H

Harlan,

I may be wrong but I believe that "synchronization distance" is the 
criteria for selecting the synchronization source from the selection 
set.  Since synchronization distance is one half the round trip delay 
plus the dispersion, it is most unlikely that servers with delays of 100 
or 200 milliseconds will ever be selected unless there is nothing else 
available.

It's generally possible to find servers with delays less than twenty 
milliseconds.  It should almost always be possible to find servers with 
delays less than 50 milliseconds.

_______________________________________________
questions mailing list
[email protected]
https://lists.ntp.isc.org/mailman/listinfo/questions

Reply via email to