Continental Aggregation

[Yakov Rekhter]

Thomas,

> > Some would argue that it is not broken, and simply replacing
> > v4 with v6 will simply introduce a probelm in this regard
> > (or amplify the problem, if you believe one exists).
> 
> And others would argue differently. For instance:

Of course... 
  
> 1) IPv4's current state of poor aggregability has a lot to do with
> legacy addresses that were handed out long ago before the need for
> CIDR-like aggregability was well understood. IPv6 starts with a clean
> slate, so addresses can be allocated more sanely from the start.
> 
> How many prefixes would be in the DFZ today if we could renumber every
> site? 

I see - your solution to site renumbering is to transition to IPv6.
Certainly one possible option. But certainly not the only one possible.
And unlikely to be most cost effective either. However, one could
certainly use a nuke to kill a fly... :-)

> 3) Some members of this list argue that *is* possible to renumber
> sites. 

This is not the matter of arguments - it is a matter of fact.

> In fact, the continued success of IPv4 appears to depend on
> this.  IPv6 is likely to make it even easier (though still not
> trivial) for end sites to renumber than in IPv4. This means v6 may be
> better able to maintain aggregability into the future.

Explain to me how IPv6 would make it *noticeably* easier, as compared
to a host running Win95 (or for that matter OS/2 that as far as I
know implements DHCP and Dynamic DNS Updates).

Yakov.