Subject: Re: Peculiar ICMP6 redirect rejection
To: der Mouse <mouse@Rodents.Montreal.QC.CA>
From: Robert Elz <kre@munnari.OZ.AU>
List: tech-net
Date: 08/19/2002 20:48:31
    Date:        Mon, 19 Aug 2002 13:20:53 +0200 (CEST)
    From:        der Mouse <mouse@Rodents.Montreal.QC.CA>
    Message-ID:  <200208191151.HAA13287@Sparkle.Rodents.Montreal.QC.CA>

  | but why does ND want to
  | install such a route to begin with?  R has no business doing ND for L
  | at all, since L isn't on the /112 R's on.

You're thinking in IPv4 paradigms.   IPv6 knows that there can be multiple
prefixes on one physical link, and that requiring bouncing packets through
a router to make that work is dumb, so has the ability to directly inform
a node that some other node with what looks to be an off-link address, is
actually on-link and can be reached directly.

  | To continue with a response to the message (those not interested in
  | watching kre and me chatter at one another can stop reading now :-),

That is probably everyone, and after this message, I think me included...

  | Because until the DNS infrastructure supports it well, renumbering
  | isn't going to become something that's undertaken casually.

Sure, but what is required to make renumbering work wasn't part of the
original discussion.   No question there are lots of things - but that
doesn't mean that each individual one shouldn't be worked toward on its
own (that is, no "renumbering without DNS support is hard, so because
of that I won't care if I also make it hard by using globals for routes")

  | That would make sense.  It also seems to at least somewhat work that
  | way.  I just now tried it.

At the minute I have nothing I can use to test this, but I will try to
remember to do it later where I also have a system with a currently
unoccupied interface.

And from your reply to Itojun ...

  | As for nonstandard netmasks, I'm not sure what's "nonstandard" about
  | any of the netmasks I'm using,

The conventional IPv6 wisdom in some quarters is that all links are
required to use a /64 (nothing bigger, nothing smaller).   There is
currently (or has just been) a debate about that in the ipv6 community.

One would hope that it might change, given that nothing enforces it,
there's no point, and lots of people want to do otherwise (want to guess
which side of that debate I take...) but I am not all that hopeful.
The people who support retaining the /64 have more power...

  | If you mean that IPv6 doesn't
  | support a /112 or /128 netmask in some sense, then I have to ask, why
  | on earth not??  What does it break, and what should I do instead?

Good questions.   The answer to the last would be /64 (according  to
the theories).  But for the first two, no reason I know of, and nothing
I'm aware of.

I doubt the netmask is related to any of the questions here.   Using
multiple prefixes on a link, without advertising them as being on link
may be stressing the implementation, but it certainly isn't non-standard.