[ntp:questions] Re: Sufficient # servers to sync to

David Schwartz davids at webmaster.com
Sat Mar 19 21:04:05 UTC 2005

"Richard B. Gilbert" <rgilbert88 at comcast.net> wrote in message 
news:IpidnYQnM441gKHfRVn-1w at comcast.com...

> Ummm....   When is the last time you saw two servers reporting exactly the 
> same time?   The difference is usually in the range 0.5 to 40 milliseconds 
> but it's there.  If they do happen to agree, is it random chance or do 
> they both actually have the correct time?

    I posted elsewhere with more detail. Servers respond with both a time 
and an accuracy estimate. A server is working if the true UTC time is within 
the time plus or minus the accuracy estimate. If you poll two servers, and 
they are both working, their times plus their accuracy intervals plus the 
round trip time *must* overlap, or else the servers are not working.

    The assumption is that we have three servers, two working, two not. 
Since two of the servers are working, by definition their time plus 
uncertainty intervals must have some small amount of overlap in which the 
true time lives. Since the third server is by assumption not working, the 
true time must lay outside its interval or the interval must be absurdly 

    In this situation, it is trivial to isolate the incorrect server from 
the two correct ones. However, one is not left with a very strong assurance 
that the other two are in fact correct.


More information about the questions mailing list