[ntp:questions] Effects of server processing time variations...

Danny Mayer mayer at gis.net
Fri Aug 5 03:36:55 UTC 2005


Som Naderi wrote:
> Group,
> 
>  
> 
> Hello all. I have a question:
> 
>  
> 
> What effects do server processing time variations have on NTP's ability to
> maintain time synchronization?
> 
>  
> 
> If the processing time between T2 and T3 (server receives packet [T2],
> processes, server sends packet [T3]) varies, let's say, randomly up to a
> second, how will this negatively impact NTP?
> 

It's a basic assumption of NTP that that will be true. The challenge has 
been to figure out the best way of dealing with such variations in such 
a way that it will quickly dampen fluctuations in the estimates (wildly 
changing clocks are unacceptible and useless). The resulting estimate is 
just that, an estimate but one that is relatively stable.

>  
> 
> As far as I understand, ideally speaking (T1-T4 RTT equations) there
> shouldn't be any impact.  I'm interested in hearing even the minutiae of
> issues as it pertains to NTP and the algorithms it uses.
> 
For that you really need to look at Dave's documentation and research 
papers. This kind of stuff has been the subject of many theses. Just 
note that Dave is continually modifying the algorithms used to improve 
stability, reduce jitter, etc. It takes a lot of analysis and design.

Danny
>  
> 
> Thanks for the help!
> 
>  
> 
> - Som
> 
>  
> 
>  
> 
> Som Naderi
> SPAWAR Systems Center, San Diego
> GPS/Navigation Division, Code 2311
> Phone: 619-553-8781
> 
>  
> 
> _______________________________________________
> questions mailing list
> questions at lists.ntp.isc.org
> https://lists.ntp.isc.org/mailman/listinfo/questions
> 




More information about the questions mailing list