On 05/23/12 10:21, unruh wrote:
On 2012-05-22, Eric S. Raymond<[email protected]> wrote:
It's possible that Kalman filtering could be useful for cleaning noise
from an NTP server's measurements of propagation delay. It's a general
technique used for all kinds of noisy time series.
Could be. Ntpd's handling of noise has always been primative. Mills
would probably claim that it is thus for robustness, but I think the
evidence is sparse. chrony uses least squares fitting to try to
eliminate the random noise, and tends to do much better than ntpd at
keeping the time near the true time ( about 2-3 times better in my
tests). It does this by retaining more information about the past
behaviour, rather than simply retaining one number or two (current time
and rate).
Note that reading that Wikipedia entry, the current ntpd scheme is a
Kalman filter it would seem.
If you are in a situation of limited memory, so that saving say the last
60 measurements is far too expensive, then a markovian model is good. We
are not in that state with modern computers.
I wish I knew more about Statistics. Clearly the behavior of individual
poll data over time tells you something about the drift rate. And
adjusting for drift over time means that they also tell you something
about the offset. And just as clearly for any set of poll data, the
smaller the delay the less the maximum offset error. I wonder if this is
a solved problem in Stats? There must be an optimum solution to making
this calculation, one that does not blindly throw away a lot of data
just because it isn't the best data of the set, nor ignore the fact that
some of the data has a better error tolerance than the rest.
--
blu
Always code as if the guy who ends up maintaining your code will be a
violent psychopath who knows where you live. - Martin Golding
-----------------------------------------------------------------------|
Brian Utterback - Solaris RPE, Oracle Corporation.
Ph:603-262-3916, Em:[email protected]
_______________________________________________
questions mailing list
[email protected]
http://lists.ntp.org/listinfo/questions
