[LEAPSECS] ISO Influence

[Poul-Henning Kamp]

In message <4D0DBBC5.9030407 at rubidium.dyndns.org>, Magnus Danielson writes:

>On 12/19/2010 03:43 AM, Warner Losh wrote:



>If using some method of alignment, such as NTP. Many computers is way off.


I could talk about this for days, because I made the mistake of letting
NTPns report such stuff.

This is from gps.dix.dk:

	NTPns > show ntpv4 0 partner
	Max partners:           10000
			   total   ours others
	partners            2338   1130   1208
	partners good       1982   1083    899
	partners bad         356     47    309
	partners > 1s        194      3    191
	partners < 1s        184     50    134
	partners < 100ms     265     44    221
	partners < 10ms      606    249    357
	partners < 1ms      1089    784    305

(Good & bad refers to the NTP standards 128msec limit.)

Obviously, people like this are beyond hope:

194.239.75.253     123 no   4 3 11 10 -17       -4916.26 [LOCALCLOCK#0]
195.14.168.226     123 no   4 3 11 10 -17       -3909.31 [LOCALCLOCK#0]
80.167.245.254     123 no   4 3 10  6 -24        4942.26 [LOCALCLOCK#1]

While others just makes the mind boggle:

192.38.7.240       123 no   3 4  1  6 -29        1955.65 [GPS]

A lot of people use NTP servers to far away, and suffer assymetric
routing, for instance Bell Canada seems to recommend my NTP server
in Denmark:

69.159.196.173     123 unk  3 3  0  6 -16   -0.352841836 [192.38.7.240] = ours
69.159.226.196     123 unk  3 3  0  6 -16   -0.348788476 [192.38.7.240] = ours
69.159.200.112     123 unk  3 3  0  6 -16   -0.345559881 [192.38.7.240] = ours
70.25.56.213       123 unk  3 3  0  6 -16   -0.344285316 [192.38.7.240] = ours
70.52.194.139      123 unk  3 3  0  6 -16   -0.341573635 [192.38.7.240] = ours
174.94.129.44      123 unk  3 3  0  6 -16   -0.341489601 [192.38.7.240] = ours
70.28.74.13        123 unk  3 3  0  6 -16   -0.340902327 [192.38.7.240] = ours
74.14.224.184      123 unk  3 3  0  6 -16   -0.340652557 [192.38.7.240] = ours
70.28.56.27        123 unk  3 3  0  6 -16   -0.339923340 [192.38.7.240] = ours
70.25.54.30        123 unk  3 3  0  6 -16   -0.339598126 [192.38.7.240] = ours
70.28.62.58        123 unk  3 3  0  6 -16   -0.339449289 [192.38.7.240] = ours
70.28.62.55        123 unk  3 3  0  6 -16   -0.339342862 [192.38.7.240] = ours

But overall, a surprising large fraction of the internet is within
a second of UTC, all uncertainties considered.


>So... after my little epidemiological analysis, which I am sure is in 

>the back of the heads of several people here, we can conclude that most 

>likely will most computers need fixes to support leap seconds, but they 

>would not experience any trouble with dropping the insertions of leap 

>seconds.


Exactly.  But these computers are just magically wished away by the
astronomers.

Poul-Henning

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.