[126215] in North American Network Operators' Group
Re: DNS for RFC3180 GLOP reverse zone ?
daemon@ATHENA.MIT.EDU (James Hess)
Thu May 6 23:15:11 2010
In-Reply-To: <20100506181242.GA23807@hedwig.net.cmu.edu>
Date: Thu, 6 May 2010 22:14:34 -0500
From: James Hess <mysidia@gmail.com>
To: "L. Gabriel Somlo" <gsomlo@gmail.com>
Cc: nanog@nanog.org
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
On Thu, May 6, 2010 at 1:12 PM, L. Gabriel Somlo <gsomlo@gmail.com> wrote: ..
> I wonder if DNS for GLOP/RFC3180 is still expected to work/be supported,
> or should I just give up :) > Thanks,
I am not sure, but I believe as a best practice, RFC3180 is
considered basically defunct at this point, it's obvious that at least
the RDNS is neglected. The problem is that it relied on mapping bits
from the AS number into the IP address bitspace.
Now that AS numbers have been extended to 4 bytes in length, and RIRs
are even about to stop differentiating between them when allocating
AS numbers, or allowing anyone to request and be sure of getting a new
16-bit ASN.
It seems that it will be impossible for the scheme to be followed in IPv4.
A more sensible BCP at this point would be to designate the entire
223/8 to IRRs, like was suggested by the BCP for 64512 -- 65535,
since most ASNs are not using GLOP addressing.
Mapping ASN bits onto multicast IP ranges is convenient but wasteful
too, once you consider >2^16 ASNs.
--
-J
