I think another way of looking at this is to consider the "domain of
reliability".

One of the advantages of Pekka's (auto)configured model for globally
unique site local addressing is that it doesn't make absolute guarantees
of global uniqueness. While the chance of globally unique site-local
addressing collision may be low (potentially non-existent), it does
still exist.

As soon as you mention "not-reliable for external communications" to
people, they are likely to use the better alternative - the global
address space they have been assigned by their provider. I don't think
many end-users would bother to try to convince their provider to route
their "unreliable on a global domain" globally unique site local address
space.

Michel, maybe my mind isn't lateral enough, but I can't think of an
example of anybody who would want to pay for guaranteed globally unique
site local addresses. Usually people seem to be happy with "good enough
and free" verses "perfect and not-free". Are there any examples you
might be able to give ?


Thanks,
Mark.

On Fri, 2002-11-22 at 15:03, Michel Py wrote:
> Brian,
> 
> > Brian Zill wrote:
> > We need to be careful here. In our rush to eliminate
> > the bad effects of the ambiguity present in site-local
> > addresses today, let's not forget that there are some
> > major plusses to existing site-local addresses that
> > are the result of this ambiguity:
> 
> I agree with most of your points, and I have made the same arguments
> myself earlier. I think though that there are ways we can have
> reasonable guarantees that a globally unique address would *not* be
> globally routable and I will repost soon what Bob Hinden and myself have
> contributed to the topic.
> 
> > 1. They're free.
> 
> A small fee would not be a show stopper for people that want true
> uniqueness.
> 
> > 2. They can be (auto)configured without having
>      to co-ordinate with some outside entity.
> 
> We can reserve a range of FEC0::/10 for these purposes, some kind of a
> hash that would produce an almost unique address. This would be the
> preferred choice of people that don't want to bother with registration
> or that want the address for free.
> 
> > 3. They cannot be externally routed
> 
> Stay tuned.
> 
> Michel.
> 
> 
> --------------------------------------------------------------------
> 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]
> --------------------------------------------------------------------



--------------------------------------------------------------------
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]
--------------------------------------------------------------------

Reply via email to