[ntp:questions] NTP Broadcast/Multicast Benefits

Brad Knowles brad at stop.mail-abuse.org
Tue Aug 2 21:54:05 UTC 2005


At 12:41 PM -0500 2005-08-02, Matt Kinard wrote:

>  Is there a threshold that one doesn't want to cross?

	Sure.  You don't want to overload your time server.  You don't 
want to have poor time sync across your network, because your time 
servers are overloaded.

>                                                        How many clients
>  is too many for one server?

	However many is "too many" for you.

>                               How does one judge?

	When you decide/discover that you've got poor time sync across 
your network.

>                                                    I guess that these
>  are the questions that I've got.

	I'm not sure anyone has any hard and fast answers for you.  A lot 
depends on the CPU you're running your timeservers on, the network 
devices, the OS, the drivers, whether or not you've got refclocks 
that you can attach directly to these servers or if they have to get 
their time from upstream servers, etc....

	You'd think that the biggest, fastest, honkingest symmetric 
multiprocessor box would be the best thing for the job.  You'd be 
wrong.  Even experienced "chime heads" have made that mistake.


	I'd suggest Googling for "Poul-Henning Kamp" and "soekris" on 
this newsgroup/mailing list.  He's had some good things to say with 
regards to building high-quality time servers.

	As for the rest, I certainly don't have any hard and fast answers 
for you, and I'm not sure anyone else does, either.  If you're lucky, 
maybe Dr. Mills will pick up on this thread and give you the benefit 
of his experience.

-- 
Brad Knowles, <brad at stop.mail-abuse.org>

"Those who would give up essential Liberty, to purchase a little
temporary Safety, deserve neither Liberty nor Safety."

     -- Benjamin Franklin (1706-1790), reply of the Pennsylvania
     Assembly to the Governor, November 11, 1755

   SAGE member since 1995.  See <http://www.sage.org/> for more info.



More information about the questions mailing list