[ntp:questions] More Granularity in the US in the NTP Pool

Richard B. Gilbert rgilbert88 at comcast.net
Fri Sep 14 00:28:09 UTC 2007


Harlan Stenn wrote:
>>>>In article <46E97DDD.6050404 at comcast.net>, "Richard B. Gilbert" <rgilbert88 at comcast.net> writes:
>>>
> 
> Richard> The quality of time and therefore the "best" servers depends
> Richard> strongly on the length of the network path between server and
> Richard> client.
> 
> Why are you discounting traffic volume?  I'm not saying you are wrong to not
> bring it up, but I wonder about it.
>

I'm not discounting traffic volume.  I have no means to measure it and, 
without some data, I hesitate to comment.  It seems to me that there are 
at least two ways in which traffic volume could affect NTP performance:
unrelated traffic on the network, and the traffic handled by a 
particular server.  I have seen what seems to be just generic traffic on 
the network bollix up NTP performance.  Many internet servers seem to 
work a LOT better between 2200 and 0700 EST when the net seems to quiet 
down.

With a broad band connection to the internet (cable, ADSL, T1, etc.) it 
would take a LOT of traffic to affect NTP very much.  Of course a few 
thousand misbehaving clients can generate an awful lot of traffic!

> And as I understand it KOD doesn't help much, especially for miscreant
> clients.

I wouldn't know about KOD! I can't very well operate a public server: my 
contract with Comcast prohibits it, and Comcast can and does change my 
IP address without notice (not recently but it has happened many times 
in the past).




More information about the questions mailing list