[ntp:questions] Re: Time mysteriously advancing with FreeBSD 5.3 and ntpd 4.2.0-a

Bjorn Gabrielsson bg at lysator.liu.se
Sat Jan 1 15:36:25 UTC 2005

Mxsmanic <mxsmanic at hotmail.com> writes:

> Richard B. Gilbert writes:
> > I normally see 
> > offsets in the 100us to 1ms range from my GPS reference clock and 
> > offsets in the 5 to to 20ms range for the best of the network servers I 
> > use.
> This is something I also don't understand very well: If you have a very
> accurate local reference clock, what's the utility of querying distant
> NTP servers?  Is it just a sanity check, or what?  Won't the local clock
> _always_ be the final reference used by NTP?
> I've been thinking of getting a DCF77 clock for my machine, if budget
> ever permits.  I suppose that would be at least as accurate as network
> synchronization (?).  A GPS clock isn't an option because I have no easy
> way to give it a clear view of the sky in my building.

Jonathans DCF-receiver should not be a budget problem. I have been
running it on Linux. With initial calibration done, I get offsets
around or below 0.5ms. Correct mounting -- away from interference
sources and aligning the antenna well -- is important. When done
correctly I will still loose reception a few times a day. I live in
Sweden (around N58E18) which is quite far from the DCF transmitter.

However the DCF will not beat the best public NTP servers. The best
external servers usually have offsets below 0.2 ms. Distance/delay to
the servers are in the 10ms to 15ms range. It seems the NTP server
infrastructure is exceptionally good here in Sweden. 

Comparing different timesources will obviously depend on your local
situation -- distance to radio time transmitters, ability to place a
GPS antenna with good sky view, quality of NTP servers in your region,
quality of network path to said NTP servers, etc. There is no "one"
solution that fits everyones requirements, location or budget!

More information about the questions mailing list