[e2e] Accuracy achievable with NTP

[David G. Andersen]

On Thu, Apr 10, 2003 at 12:16:04PM -0400, Woojune Kim quacked:
> Hi,
> 
> I've been looking for info on how accurate NTP can be. I've looked in various NTP related sites and still not quite sure.
> 
> Say I have a LAN like environment with delays in the range of 10s of ms. How accurate a time can I get from NTP, assuming I have the appropriate HW clocks on my host ?

  Worst case is 1/2 the delay on your network.  What kind of LAN do
you have that gets 10ms delay?  Yuch. :)

  In reality, you can usually get w/in a few miliseconds on a decent LAN.

> Say I have a corporate network like environment, relatively segregated from the Internet at large, with delays in the range of 100ms. How accurate a time can I get ? 
> 
> Say the delay range goes upto 500ms, how accurate can the time be ? 

   The answer to this depends entirely on the variance of your link
degree, and the degree to which the link delays are asymmetric.  If all
delays are perfectly symmetric and there's little variance, in theory,
you can get the same couple of miliseconds accuracy.  NTP's algorithms
assume symmetric delay when factoring out the RTT.  Or, if you knew
the one-way delays involved, you could correct manually.

  In an "arbitrary" network topology, you could be looking at errors of
up to a few hundred miliseconds.  But with realistic topologies and
no gross asymmetry, probably 10ms accuracy.

  There's a lot of data on ntp.org and David Mills' website.  Also,
these questions come up a lot in comp.protocols.ntp;  a google newsgroup
search would be productive. 

  -Dave

-- 
work: dga at lcs.mit.edu                          me:  dga at pobox.com
      MIT Laboratory for Computer Science           http://www.angio.net/
      I do not accept unsolicited commercial email.  Do not spam me.