Date: Tue, 12 Mar 2002 11:20:06 +0200
From: Markku Savela <[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
| My opinion on this is: these ID related issues all belong to
| IP-over-foo documents. The rfc-2373 should only talk about (prefix=n,
| interface ID = 128-n), and not care at all what the ID bits are.
I agree.
| Thus, all references to EUI-64 should be removed from 2373.
Yes. Exactly.
2373 should go no further than recommending that IPv6 over foo specs
should use no more than 64 bits as the interface ID. The language
that does that can be strong - but no matter how strong it can never
be more than a recommendation, the authors of a later standard are always
free to override an older one (assuming they can get rough consensus, ...)
It needs to be made very very clear that no-one can ever take the low 64 bits
of an IPv6 address, and attempt to make any sense out of them at all.
The bits have meaning only within the scope of the prefix that assigns
them, and can be interpreted only by knowing the assignment rules of
whoever controls the prefix.
Common sense rules like "set (the inverted) u=1 when autoconfiguring"
make sense - they make it easy to mix & match assignment styles within
a prefix - but no-one other than the owner of the prefix in question
should be relying upon such a rule being applied.
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]
--------------------------------------------------------------------
