This is output from my 17 year old server, i386, 32-bit, no TXCO, cheap
hardware, sitting in an airconditioned office where staff keep fiddling
with the thermostat (the large unit is 20 years old, so you are either cold
or very cold).
root@ntpmon:~# uptime
10:20:02 up 14 days, 9:33, 1 user, load average: 0.04, 0.03, 0.00
root@ntpmon:~# ntpq -pnu
remote refid st t when poll
reach delay offset jitter
=======================================================================================================
*SHM(1) .PPS. 0 l 93 64
376 0ns -5.526us 5.343us
-SHM(0) .GPS. 0 l 27 64
377 0ns 74.723ms 63.766ms
-2604:a880:1:20::17:5001 64.142.122.37 2 8 6 64
377 190.57ms 964.59us 157.85us
-2a03:b0c0:1:d0::1f9:f001 193.62.22.74 2 8 54 64
377 229.74ms 2.4097ms 177.81us
-2a00:d78:0:712:94:198:159:11 192.87.36.4 2 8 32 64
377 195.16ms 13.216ms 117.49us
-2606:4700:f1::1 10.23.8.4 3 8 53 64
377 38.222ms -2.133ms 199.36us
+2403:5000:171:11::2 .MRS. 1 u 36 64
377 38.352ms 591.67us 71.542us
2404:e800:3:300:218:186:3:36 .STEP. 16 u - 1024
0 0ns 0ns 477ns
+118.143.17.82 .MRS. 1 u 59 64
377 36.238ms -1.341ms 26.321us
-193.204.114.231 .CTD. 1 u 54 64
357 255.54ms 40.967ms 63.054us
131.107.13.100 .STEP. 16 u - 1024
0 0ns 0ns 477ns
root@ntpmon:~#
2604:a880:1:20::17:5001 is ntp1.glypnod.com. 2403:5000:171:11::2 is in
Hong Kong.
If I can get this on crap hardware, poorly maintained and tuned, on the
other end of the world from glypnod.com , I assume there is low-hanging
fruit still possible. Letting Hal or Gary near it should push it to 100us,
I assume.
PS: My official reason for not using newer hardware is "I am making sure
NTPsec and gpsd do not drop support for i386". :-)
--
Sanjeev Gupta
+65 98551208 http://www.linkedin.com/in/ghane
_______________________________________________
devel mailing list
[email protected]
http://lists.ntpsec.org/mailman/listinfo/devel
