To:
Robert Elz <kre@munnari.OZ.AU>
CC:
Jun-ichiro itojun Hagino <itojun@iijlab.net>, Pekka Savola <pekkas@netcore.fi>, dnsop@cafax.se
From:
Masataka Ohta <mohta@necom830.hpcl.titech.ac.jp>
Date:
Fri, 22 Nov 2002 23:46:00 +0859 ()
In-Reply-To:
<12624.1037937786@munnari.OZ.AU> from Robert Elz at "Nov 22, 200203:03:06 pm"
Sender:
owner-dnsop@cafax.se
Subject:
Re: comments on dnsop-ipv6-dns-issues-00
kre; > | Your assumption that source address of a incoming packet is reachable is > | improper. > > The source is more likely reachable than some DNS server at some unknown > location There are multiple DNS servers. Moreover, you must, anyway, rely on DNS for forward lookup. > - at least we know the source address is actually alive and connected > at the time. Source host is, source address of the host may not. > Personally, I have lost essentially all faith in the usefulness of addr->name > translations at all, and wouldn't mind simply saying "cannot be done" (in > general of course, it might sometimes work, just don't depend upon it). But > if it must be done, the node info query way is good enough. The primary benefit of having a reverse tree is that the tree makes address assignement unique. > | Never assume routing symmetric. > > What does that have to do with anything? v6 people including itojun persistently assume it, for example, for such topics as a source address selection. Masataka Ohta #---------------------------------------------------------------------- # To unsubscribe, send a message to <dnsop-request@cafax.se>.