Re: [chrony-dev] [GIT] chrony/chrony.git branch, master, updated. 1.25-pre1-18-g20a4340

[Bill Unruh <unruh@xxxxxxxxxxxxxx>]

On Wed, 13 Apr 2011, Miroslav Lichvar wrote:

> On Wed, Apr 13, 2011 at 09:48:46AM -0700, Bill Unruh wrote:
> > On Wed, 13 Apr 2011, Miroslav Lichvar wrote:
> > > sd = sqrt(inst->variance)
> > > sd_weight = 1.0 + SD_TO_DIST_RATIO * (peer_distances[i] - min_distance) / sd;
> > > weights[i] = sd_weight * sd_weight;
> >
> > That use of sd in the weights is new in the code. It used to be be that the weighting was
> > purely by the ratio of the excess distance over the min_distance. Not sure
> > why the sd is in there instead now. Did you add it?
>
> Yes, I did. The idea behind it is that adding a constant to the
> distance shouldn't change the weights. Two servers with equal
> jitters, but different delays should give the same results.

I am not sure I understand that. The old one basically said that until the
distance was about 2 the mindistance the weights did not change. Ie, it
downplayed minor fluctuations in the distance. Adding a constant to the
distance (assuming this did not change the min_distance) still changes the
weights in yours (and a lot if sd gets small) . If you wanted you could put in 
sd+min_distance as the divisor. But this seems the wrong thing to do anyway,
because really the importance of small differences in the distance should
always remain small. This weighting is based on the distance (It is the
equivalent of the "throw out 7 out of 8 measurement" that ntpd does ( which
seems a terrible waste of precious data to me). Also, the normalisation of the
weights should not matter. If you wanted, you could try doing

mindist/sd (1+ SD_TO_DISR_RATIO*(peer_distance[i]-min_distance)/min_dist) 
which would maintain the fact that small changes of distance are unimportant.

Also, if a server is further away, it IS less reliable. There could be a
constant offset-- outgoing times being a set amount more than incoming. Ie, if
a server has a round trip of 1 sec, and another has a roundtrip of 1ms, even
with the same jitter I would trust the second more.

>

--
William G. Unruh   |  Canadian Institute for|     Tel: +1(604)822-3273
Physics&Astronomy  |     Advanced Research  |     Fax: +1(604)822-5324
UBC, Vancouver,BC  |   Program in Cosmology |     unruh@xxxxxxxxxxxxxx
Canada V6T 1Z1     |      and Gravity       |  www.theory.physics.ubc.ca/

---
To unsubscribe email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject.
For help email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.