[ntp:questions] NTP tunning for OWD measurements

David Woolley david at ex.djwhome.demon.invalid
Fri Oct 26 16:32:55 UTC 2012


Brian Utterback wrote:

> 
> You may not want to, but that is the way to do it. If all of your 
> systems are local, you could simply use a single PPS source and fan it 
> out to all of the systems.  If not, then you should have a local source 
> to each system. Otherwise, you should do just as you have done, and then 
> look at the spread of time offsets between each of the four stratum one 
> servers and the round trip times. That gives you the maximum error in 
> your measurements.

He needs to use root delay as the error estimate in this case, because 
it is a given of the problem that round trip times are not symmetric.

Basically, unless he can constrain the delay asymmetry to his NTP 
servers, the timing error will depend on the parameter he is trying to 
measure.



More information about the questions mailing list