[time-nuts] Divide by five

Paul tic-toc at bodosom.net
Sun Nov 9 16:29:44 EST 2014


On Sun, Nov 9, 2014 at 3:44 PM, Poul-Henning Kamp <phk at phk.freebsd.dk>
wrote:

> NTP does indeed find the best clock from the subset of clocks
> which pass its sanity check, and then it uses only that one.


Normally I wouldn't question your NTP assertions but I suspect a great many
people believe   "NTPv4 Algorithms Specification" or the 18-Jul-2012
document on Mitigation Rules.

>From <http://www.eecis.udel.edu/~mills/ntp/html/prefer.html#miti>
"2. Combine Algorithm

The clock combine algorithm uses the survivor list to produce a weighted
average of both offset and jitter. Absent other considerations discussed
later, the *combined offset* is used to discipline the system clock, while
the *combined jitter* is augmented with other components to produce the
system jitter statistic inherited by dependent clients, if any.

The clock combine algorithm uses a weight factor for each survivor equal to
the reciprocal of the root distance. This is normalized so that the sum of
the reciprocals is equal to unity. This design favors the survivors at the
smallest root distance and thus the smallest maximum error."

Or more succintly (from a Mills ppt) Θ=combine(θj) where Theta is the
(estimated) system clock offset and the theta_j are the survivor offsets.

You're saying these documents affirming clock combining are wrong?  Or have
I gone terribly astray?


More information about the time-nuts mailing list