[ntp:questions] Re: NTP adjustment mechanism

Drk Ryan ryandrk at hotmail.com
Mon Sep 1 20:40:06 UTC 2003


Thanks for the reply Ulrich. Sorry if my question seemed a bit vague.
I can understand that the values I mention would not be exact (due to
'fine wobble').
But is the mechanism I describe essentially how NTP adjustments
happen?

Drk 

Ulrich Windl <Ulrich.Windl at RZ.Uni-Regensburg.DE> wrote in message news:<m3smngissp.fsf at pc5234.rz.uni-regensburg.de>...
> Depending on the scale you are looking at it, I guess every clock does
> a bit of fine wobble; so does the NTP clock model. However it's not as
> cruse as you do describe (you may read the clock during that 0.25
> seconds to verify). You also should be more specific with your questions...
> 
> Regards,
> Ulrich
> 
> ryandrk at hotmail.com (Drk Ryan) writes:
> 
> > If clock adjustment is limited by tickadj/ticksize eg 4/10000 = +-400
> > ppm , will this essentially mean that for a +100 ppm clock drift, NTP
> > will run at -400 ppm for the first 0.25 of each second making
> > adjustments each tick => accumulate an error (relative to actual
> > clock) of -100 usec and then over that second the actual clock will
> > gain +100 => balance out ..=> this means that worst case scenario NTP
> > error will be at 0.25 sec in each sec interval and will be -100 (due
> > to adj mechanism) +25 (accumulated error at thsi stage due to inherent
> > error) = -75 msec...is this correct?



More information about the questions mailing list