[128824] in North American Network Operators' Group
Re: Numbering nameservers and resolvers
daemon@ATHENA.MIT.EDU (John Kristoff)
Tue Aug 17 11:48:28 2010
Date: Tue, 17 Aug 2010 10:48:21 -0500
From: John Kristoff <jtk@cymru.com>
To: Sven Olaf Kamphuis <sven@cb3rob.net>
In-Reply-To: <Pine.LNX.4.64.1008171201090.26804@a84-22-97-10.cb3rob.net>
Cc: nanog@nanog.org
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
On Tue, 17 Aug 2010 12:11:56 +0000 (UTC)
Sven Olaf Kamphuis <sven@cb3rob.net> wrote:
> nowadays, i'd simply put them all on the same /24 which you simply
> announce on different pops
I would raise a red flag of caution with this approach especially for
services that need to be reachable outside your network If there
is a a snafu with said /24 prefix, particularly outside your own routing
domain, a reachability problem could persist for an extended period
and you'd be in a difficult position to solve it on your own.
For instance, if it flaps and someone, for better or worse, dampens that
route, that could mean an extended outage for all those hosts until the
damping period timer expires.
On a related note, some systems and folks have taken multiple unique
origin ASNs as a measure of diversity. In pratice, unless there is
some odd AS path mangling going on for your specific routes, which is
unlikely, one can properly instrument diversity using a single origin
ASN with multiple prefixes. Its the path and the prefix that matters,
much less the ASN.
John
