[ntp:questions] Re: Philosophical question about strata

Danny Mayer mayer at ntp.isc.org
Fri Oct 14 01:30:21 UTC 2005

Lee Sailer wrote:
> First, thanks to everybody for replying.  However, my original question
> was not as clear as it could have been.  Let me continue...
> My company has 6 GPS time sources at various worldwide locations.  They
> are stratum 1.
> Every once in awhile, one or the other of them has a harddware problem.
>  As a result, that one will drift a little.  That is, its offset will
> be 1, 2, or more seconds different from the other five.  When we detect
> this, we replce the hardware as a matter of course.
> Now, it seems to me that a stratum 2 server will never have this
> problem.  It gets time from several places, and uses the various
> algorithms to try to make sense.  The stratum 1 gets time from a bunch
> of satellites, and then if it screws up the calculation or has some
> sort of hardware glitch, then it is screwed.
> Therefore, I conclude:
> 1.  If everything is working, stratum 1 is better than stratum 2.
> 2.  If the stratum 1 hardware is suffering a slow degradation
> breakdown, then the stratum 2 will be better.
> Whaddya think 8-)

It's a good question, but no it isn't. If you are using stratum 1 or any 
other stratum servers then you need to use 4-5 of them. If your system 
sees that one of them is drifting away from the rest of the group it 
will be dropped from consideration as a candidate since the other 
servers are still agreed on the expected value and it's falls out of the 
delta required for consideration. The NTP algorithms are quite punative 
of this. It doesn't matter what the stratum is here. The stratum 1 
servers will provide better accuracy in general though network distance 
and congestion can be a major factor.

> lee

More information about the questions mailing list