nanog mailing list archives

Re: Non-GPS derived timing sources (was Re: NTp sources that work in a datacenter)


From: Sean Donelan <sean () donelan com>
Date: Mon, 2 Jun 2003 17:55:45 -0400 (EDT)


On Mon, 2 Jun 2003, Leo Bicknell wrote:
Could it be that providers actually have multiple sources, but for
some reason GPS is always picked as the primary source for the
public facing function?  At least a few providers keep their actual
sources (the receivers themselves) "hidden", and provide a unix box
syncing to all of them as the front end.  From my limited knowledge,
that front end box will only show the one source it has picked as
"best".

Like all answers, it depends.

A relatively small number of providers have blocked access to their
"master" NTP servers. So you can't see the current source from the next
stratum down.  You get a domain name, like ntp-1.ispdom.ain, for the
source.

But a large number permit queries to their NTP servers. They will report
all the stratum 0 sources available, including the current sync'd source.
Almost no providers use more than one stratum 0 source per NTP server.
The people with lots of stratum 0 sources are almost never network service
providers. Generally NSPs have 1 stratum 0 source, a two or three stratum
1 peers, and lots of stratum 2s.  Most providers don't have a dedicated NTP
infrastructure, so their tickers are running on routers and general
purpose servers.

Due to the way NTP works, it is a self-directed hierarchy.  A stratum 1
chimer is only at the top while its synced to a stratum 0 source.


Current thread: