Date: Tue, 27 May 2003 13:22:23 +0200
From: Brian E Carpenter <[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
| Yes there is. As Bob Moskowitz discovered ages ago on ANX, and others have
| discovered since, you can't operate VPNs among a set of users of Net 10
| in any reasonable way,
Of course not.
| and of course the same applies to a set of users of FEC0::/10.
It applies to a set of users under FEC0::/48, yes, certainly.
Whether it applies to FEC0::/10 or not depends upon what the users
have done with the other 38 bits.
| The problem goes away with unique /48s under FC00::/8.
It would, if they were unique.
| Now, as a pragmatist, I would probably settle for a pseudo-random
| and probably-unique /48, but not everybody will.
Anyone who doesn't is deluding themselves, as that's all they're going
to get, without some method of verifying the uniqueness (some that
everyone must use for the number to be useful).
| I can just imagine a phone call in which I recommend to IBM's chief
| network architect to use address space that *probably* nobody else is using.
How are you ever going to be able to do better. Or rather, if you like,
once IBM gets such a number, you leak it to me, and then I'll guarantee
that someone else is using it...
What you're getting here is really just some kind of belief that if there
is a conflict, you have some justification for claiming that my version of
3 is the one true 3, and everyone else is a fraudulent copy (and in the case
above with IBM, mine would be, obviously, but not necessarily all others).
| Do other people think we need a guaranteed-unique mechanism
| for limited-scope addresses, or is probably-unique enough?
What people want is nice. How you actually get a a guaranteed-unique
number is another question, and one I am still looking for an answer to.
That is, without some means of ensuring the number is unique, which to
me, means making use of the uniqueness in some way (ie: non-unique numbers
don't work usefully) I can't see any method at all of actually guaranteeing
uniqueness, as anyone can, maliciously or not, simply make up any number
they like, and use it. And the number they make up might be the same
number as the number you paid for. And really, their right to make up
numbers is no less than anyone else's (including IANA, ICANN, ...).
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]
--------------------------------------------------------------------
