[ntp:questions] 500ppm - is it too small?

Unruh unruh-spam at physics.ubc.ca
Fri Aug 14 01:40:41 UTC 2009


John Hasler <john at dhh.gt.org> writes:

>Unruh writes:
>> All I am saying is that there is no rational reason for the current
>> limits.

>I had the impression that Dr. Mills' stability concerns have to do with
>the stability of networks of servers, not of individual servers.

The model is a heirachical model-- every level only depends on levels
above it, not below.Thus you cannot get freedback I thought. Also, the
primary issue here is not the changes in rate, but the value of the
absolute rate. Ie, the fact that machine A requires a 1500PPM drift
correction should not impact anything on the rest of the network, and
even changes in the drift over 1500PPm it would seem would have far less
of an effect than do steps in the time (infinite rate changes) which ntp
does allow and do (i fthe offset is  >128ms)
And in one of his comments a while ago, he stated that the 500PPM was an
arbitrary limit, which could be changed if one wanted (at least that is
what I remember). 

>-- 
>John Hasler 
>john at dhh.gt.org
>Dancing Horse Hill
>Elmwood, WI USA




More information about the questions mailing list