[28745] in North American Network Operators' Group
Re: CIDR Report
daemon@ATHENA.MIT.EDU (Peter Galbavy)
Mon May 15 06:54:31 2000
Message-ID: <000b01bfbe5b$a1910380$cfa728c3@knowledge.com>
From: "Peter Galbavy" <peter.galbavy@knowledge.com>
To: "Steven M. Bellovin" <smb@research.att.com>, <rmeyer@mhsc.com>
Cc: "'Owen DeLong'" <owen@dixon.delong.sj.ca.us>, <nanog@merit.edu>
Date: Mon, 15 May 2000 11:28:54 +0100
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Errors-To: owner-nanog-outgoing@merit.edu
> Of course, note that the real problem isn't the memory space, it's the
> CPU time to calculate the routing table updates. Note, too, that the
> more unaggregated prefixes in the net, the more changes will need to be
> propagated to every other router in the default-free zone, implying the
> need for more instances of a more expensive calculation.
Not trying to teach a mailing-list full of grandmother's to suck eggs, but
IFF all 2^32 (-1?) addresses were routed as hosts, then there is no index,
and a linear lookup cost. Stupid thing to do, but a 2^32 * 1 or 2 bytes of
linear address -> interface table.
Peter
