[ntp:questions] Tiime Sync solution
Brad Knowles
brad at stop.mail-abuse.org
Tue May 31 19:04:53 UTC 2005
At 8:16 PM +0200 2005-05-31, Adrian 'Dagurashibanipal' von Bidder wrote:
>> relatively slow ADSL line, I'm usually down in the low single-digit
>> *microsecond* region.
>
> You sure?
I was. I was quite certain that the column displays in ntpq were
measured in microseconds, but going back to the documentation at
<http://www.eecis.udel.edu/~mills/ntp/html/ntpq.html>, I see they are
actually in milliseconds. My bad.
> Hmmm. Eric: Scrap the '10ms or better is pure luck' - I see that my
> server is below 5ms offset, with hand-picked off-site servers, this is
> over a little used ADSL connection.
Agreed. Looking at my stats again, and I see that my estimated
offset is currently just over 3ms, with a stability of 0.008.
Pick some good time servers that are nearby, and you should be
able to fairly easily get down into the single-digit millisecond
range, even on lower-end hardware with less than ideal networking --
such as my several year-old PowerBook G3/400 running MacOS X 10.3.9,
running on an 802.11b wireless network which is connected to a fairly
slow ADSL line.
However, to get down to 1ms or below, you're going to have to do
a lot of hard work, and the most important parts are choosing the
right hardware platform, choosing the right OS on which to serve that
data, and choosing the right refclocks (e.g., a good GPS device which
is used to discipline a thermally stabilized oscillator).
With lots of hard work, and some luck in finding used Rubidium
clocks at reasonable prices, Poul-Henning Kamp was able to put
together some devices that achieved sub-millisecond accuracy. But
that's a lot of hard work.
--
Brad Knowles, <brad at stop.mail-abuse.org>
"Those who would give up essential Liberty, to purchase a little
temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755
SAGE member since 1995. See <http://www.sage.org/> for more info.
More information about the questions
mailing list