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:

1. They're free.
2. They can be (auto)configured without having
   to co-ordinate with some outside entity.
3. They cannot be externally routed (Some would
   consider this to be a minus as well).

I believe 1 and 2 can be solved (fairly) easily by other means.  The big
problem is number 3. The ambiguity is essential to preventing them from
ever being routed.  An IETF edict not to route some new form of
non-ambiguous addresses will be ignored by those who wish to route them.
This will likely lead to pressure to turn these new addresses into
yet-another type of global address with all the associated routing table
explosion inherent to non-aggregatable address space.  Another way to
look at it is that the current ambiguity in site-locals is a means for
protecting the global routing space.  If we do away with the ambiguity,
we need another method for keeping the global routing tables sane.

The only class of solution which I think will truly make everyone happy
is to come up with an *aggregatable* globally unique address space that
still has properties 1 and 2.  In other words, some sort of
provider-independent global address space.  Properties 1 and 2 are much
easier when the aggregatable property is not required.  The reason we're
in the state we are today is because this is a hard problem.

--Brian

--------------------------------------------------------------------
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