Date: Sun, 9 Jun 2002 14:54:33 +0300 (EEST)
From: Pekka Savola <[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
| Me neither. More probable outcome is that someone starts to request that
| people implement NATv6, because 1) they're already used to it (and like
| its "security") in v4 world,
This is bogus - but then again we see here over and over where people are
only happy when nothing changes (and we're seeing it again here).
| and 2) they think it's easier for them to do NAT than to renumber.
And they're right. And that's what site locals are good at - their
main function in my mind. If site local addresses were used for all
internal communications (whatever "internal" is defined to be), and
globals only for external communications, then renumbering IPv6 becomes
close to isomorphic to renumbering IPv4 using NAT - all the nodes need
to be renumbered, but with IPv6 that's close to automatic, so not much
of an issue. All internal communications are unaffected, as they're
using site local (just like 1918 addresses in v4 with NAT), and all external
communications are disrupted, because the global address has changed
(just like IPv4 with NAT).
| Site-locals were born in the era that not all sites had internet
| connectivity.
That was one of their planned uses, in my mind, always a minor one.
| It's
| just easier for people to use a global address block (even if we define
| that address block to be 3ffe:eff3::/32 or whatever) even with these
| "internal needs" (note: I believe there should be _something_ that does
| not require you to fill any kind of paperwork).
Huh? What difference do you think that would make? We just define
fec0::/10 as a "global" address block, and we're done according to this
theory. That's absurd - the bit pattern used can't possibly be relevant.
What matters is whether or not there is some block of addresses that can
be arbitrarily simply "taken" and assigned for local use, and which won't
work for communication that isn't local (whatever definition of local has
been chosen to apply).
If there is, then all the same issues remain - how to discover the address,
how to decide when to use it, how to prevent it leaking where it shouldn't,
etc - none of which is in the slightest impacted by the bit pattern chosen.
And on this, Steve Bellovin went on to suggest ...
| Yah. Let's pick a prefix, and tell folks to pick a random number
| (more precisely, use an RFC 1750-compatible RNG) to fill out the rest
| of the high-order bits to a /48 or a /64.
So, let the prefix be fec0::/10 and then when you have just suggested is
Paul Francis' NUSLA's - see the thread from Feb 2001 with the illumination
Subject of "Wade through the archives" (and perhaps also the threads on
"Question on scopes involving IPv6 addresses" and "what is a site??" from
March '01).
This stuff seemed to just drift into silence, without any real discussion,
which it really deserved having.
kre
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
