[ntp:questions] Bad NTP servers jeopardizing the pool.ntp.org initiative

Danny Mayer mayer at ntp.isc.org
Mon Apr 2 12:09:56 UTC 2007

Wolfgang S. Rupprecht wrote:
> mayer at ntp.isc.org (Danny Mayer) writes:
>> Which point would that be? Having NTP servers with varying addresses is
>> a good idea? 
> It appears to me that pools is on the verge of collapse from way too
> high a load.  There are only ~1000 pools servers right now.  (ref:
> http://www.pool.ntp.org/ right side bar.)  The load on each server is
> described as 5-15 packets per second for a data rate of 10-15kbit/sec.
> (ref http://www.pool.ntp.org/join.html) That comes out to ~5GByte per
> month.  I'm not sure my ISP will be that happy with me if I committed
> to that high an added load.  I would have much fewer worries if the
> load were 1/10th what it is.  To do that pools has to expand to 10x
> its current size.  Opening pools to dynamic hosts would allow a
> significantly larger pool of host to apply.
> (* (/ 15e3 8) 60 60 24 30) 4.86 GBytes / month

I understand the load issue but that's not what's being discussed here.
Trying to use something that would make the situation worse doesn't help.

>> ... there are a lot of other NTP clients out there. We already have
>> cases of people hardcoding IP addresses of NTP servers without
>> unauthorization of the owner of the NTP server.
> The fastest way to break people of the habit of wiring in IP addresses
> would be to only allow dynamic hosts into pools.  Unless they use the
> hostname as they are instructed to, they won't get any time service
> after a while.

And as I keep saying, this is a set and forget situation. They set up
their NTP service, makes sure it's working at that moment and move on
and never go back to check.


More information about the questions mailing list