[154287] in North American Network Operators' Group

home help back first fref pref prev next nref lref last post

RE: [c-nsp] NTP Servers

daemon@ATHENA.MIT.EDU (Keith Medcalf)
Sat Jun 30 18:40:11 2012

Date: Sat, 30 Jun 2012 16:39:36 -0600
In-Reply-To: <8C6A44F7-FBC7-4F42-9830-22572A96FF3C@puck.nether.net>
From: "Keith Medcalf" <kmedcalf@dessus.com>
To: "nanog@nanog.org" <nanog@nanog.org>
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org


> Or you can ask the it guys to use a windows server... Eg:
> 
> http://support.microsoft.com/kb/816042

That is a joke Jared?  You left off the smiley.

Windows doesn't do NTP out-of-the-box (Microsoft assertions to the contrary=
 notwithstanding).  You can build a reasonably working standard daemon, how=
ever don't expect time to be very accurate.  Windows out-of-the-box can kee=
p time +/- 10 minutes or so using the Microsoft lets-pretend-NTP.

You can build the current standard NTPD distribution on Windows.  You can a=
lso spend lots of time to make it "work as well as possible" (once you mana=
ge to get it to compile, that is).  Even so, when you have configured it to=
 the optimality of accuracy, this is what you can expect:

     remote           refid      st t when poll reach   delay   offset  jit=
ter
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D
+tic.nrc.ca      .PPS.            1 u   13   64  377   55.544    5.913   0.=
870
-tac.nrc.ca      .ATOM.           1 u   48   64  377   56.188    4.768   3.=
041
-toc.nrc.ca      .ATOM.           1 u    1   64  377   55.485    4.758   0.=
981
+tick.usask.ca   .GPS.            1 u   34   64  377   19.566    6.942   5.=
699
*tock.usask.ca   .GPS.            1 u   29   64  377   19.665    5.955   1.=
937
-clock.isc.org   .GPS.            1 u   37   64  377   53.091    8.311   0.=
649
+clock.sjc.he.ne .CDMA.           1 u   48   64  377   43.591    6.066   2.=
501

offset:               0.005955 s
frequency:            23.346 ppm
poll adjust:          -30
watchdog timer:       47 s

is about the best you will get.  Statistics are pretty awful:

Date        #     O.Avg     O.Median  O.Range   O.CI      O.Skew    O.Kurt =
   F.Avg     F.Median  F.Range   F.CI      F.Kurt
2012-01     899   0.765559  0.004198  20.05221  0.000371  -0.56023  0.75115=
1  21.31698  20.9705   2.9685    0.108050  -0.88068
2012-02     9673  0.237434  -7.46502  59.75607  0.000156  -1.43583  8.60908=
5  19.01126  19.3495   5.2995    0.040683  -0.54578
2012-03     1380  -0.02157  -14.8416  44.00043  0.000124  -1.08589  4.55904=
9  18.08941  16.822    7.536     0.045387  0.268831
2012-04     1322  0.196654  21.16261  106.1250  0.000141  -0.48643  26.0586=
8  17.56811  16.812    6.111     0.040561  -0.38021
2012-05     8849  0.118125  27.44213  72.01526  0.000161  0.296114  8.93942=
9  17.88685  15.2595   9.3195    0.080186  1.121740
2012-06     1457  0.409662  -20.2809  63.32684  0.000114  -1.44144  11.9823=
7  20.50724  19.5425   6.7425    0.042372  -0.08891
            6102  0.201651  21.16261  106.1250  6.065429  -0.84354  13.7816=
1  18.71838  16.1125   10.1725   0.023443  1.215941

This is from a custom ntpd build using the highest precision that it can ma=
nage to coerce from Windoze.

Of course, this may be accurate enough for most uses -- at least it does no=
t have to time-step.

Doesn't compare to ntpd on linux on an 80286 with 640K ram booting from a f=
loppy, which can maintain time sync within less than 1 ms easily.

---
()  ascii ribbon campaign against html e-mail
/\  www.asciiribbon.org





home help back first fref pref prev next nref lref last post