On Sep 16, 2007, at 11:03 AM, Chuck wrote:
> does anyone have a spec on the average duration time an active time  
> request /
> response lasts? we are trying to find connection states of 1 in / 
> proc/net/udp
> and find it almost impossible.

A typical NTP request/response happens in less (much less?) than 1ms  
if you're doing queries to localhost, and even queries to other  
timeservers on the same local switched subnet tend to have latencies  
of 0.2 - 0.3 ms.  Otherwise, the network round-trip latency is going  
to dominate.

-- 
-Chuck

PS: If you'd like to know how to measure this, try "ntpdate -d" or  
"ntpq -p":

41-ns1# ntpdate -d localhost
17 Sep 13:14:28 ntpdate[62347]: ntpdate [EMAIL PROTECTED] Fri Sep 14  
19:54:16 GMT 2007 (1)
Looking for host localhost and service ntp
host found : localhost
transmit(127.0.0.1)
receive(127.0.0.1)
transmit(127.0.0.1)
receive(127.0.0.1)
transmit(127.0.0.1)
receive(127.0.0.1)
transmit(127.0.0.1)
receive(127.0.0.1)
transmit(127.0.0.1)
server 127.0.0.1, port 123
stratum 2, precision -18, leap 00, trust 000
refid [127.0.0.1], delay 0.02580, dispersion 0.00000
transmitted 4, in filter 4
reference time:    ca9933bc.1b89fc92  Mon, Sep 17 2007 13:11:24.107
originate timestamp: ca993474.8e1878e2  Mon, Sep 17 2007 13:14:28.555
transmit timestamp:  ca993474.8e0d96db  Mon, Sep 17 2007 13:14:28.554
filter delay:  0.02583  0.02582  0.02580  0.02580
          0.00000  0.00000  0.00000  0.00000
filter offset: -0.00000 -0.00001 -0.00000 0.000007
          0.000000 0.000000 0.000000 0.000000
delay 0.02580, dispersion 0.00000
offset -0.000003

17 Sep 13:14:28 ntpdate[62347]: adjust time server 127.0.0.1 offset  
-0.000003 sec

42-ns1# ntpq -p
      remote           refid      st t when poll reach   delay    
offset  jitter
======================================================================== 
======
LOCAL(1)        .LOCL.           10 l    6   64  377    0.000     
0.000   0.004
+timekeeper.isi. .GPS.            1 u  178  512  373   96.569     
4.535   4.901
*nist1-ny.witime .ACTS.           1 u  209  512  377   14.222     
0.007   0.117
-clock.isc.org   216.200.141.8    2 u  197  512  367   85.034    
-3.226   0.200
+clock2.redhat.c .CDMA.           1 u  126  512  377   46.081    
-6.363   0.671
-pi.codefab.com  18.26.4.105      2 u  237  512  177    0.225    
-2.984   0.054
-shot.codefab.co 65.212.71.102    2 u  141  512  377    0.265    
-3.782   0.595
-pong.codefab.co 209.51.161.238   2 u   44  512  377    0.270    
-0.463   1.580

The last three machines are all configured as NTP peers, and are  
showing about a quarter of a millisecond delay...


_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers

Reply via email to