[ppml] Soliciting comments: IPv4 to IPv6 fast migration

[michael.dillon at bt.com]

> Here's the simple solution which stays strictly within ARIN's 
> purview as the resource steward:
> 
> http://bill.herrin.us/arin-policy-proposal-simple.html

More complexities...  You wrote:

   ARIN shall immediately waive the host count count 
   requirement for applicants seeking IPv6 end-user 
   assignments. All other requirements for assignment 
   remain in full effect.

Waivers are bad things to have in policy. It's like the sword of
Damocles because you know that the waiver could end at any time. Policy
should be clear and simple. It should not have a waiver in it.

Host counts do not exist in the IPv6 policy as far as I am aware. This
reference seems to refer to the requirement for allocations of /48 to
200 sites in the ISP allocation policy. Yet you refer to the end user
assignment policy which has no such requirement and nothing else that I
can see which looks like a host count.

Policies should never say that all the rest of the policy still applies.
If you can't make your special case clear in and of itself, then there
is something wrong with the whole section of policy and it needs to be
restructured and rewritten.

Please stop suggesting these overly vague and general policies. We have
some serious policy issues that need to be on the agenda in Albuquerque
if we want to make an impact on the IPv4 endgame. As Bill Woodcock
pointed out, we are not alone. There are proposals being made in 4 other
RIRs. Is there anything good there? Anything to learn there?

And let's not forget that we are not the only such number allocation
organization around. How did NANPA deal with phone number shortages.
Should we be implementing some of the usage reporting and run-out
reporting that NANPA requires? What about FEMA? They often deal with
resource shortage and allocation issues. Can we learn something from
them?

And please remember that we do *NOT* run the Internet. We do *NOT*
engineer the Internet ecosystem. We can *NOT* mandate routing and other
ISP operational issues. And we *DO* have the responsibility to allocate
IPv4 addresses for non-Internet use as well.

If you really, really, really want to help solve the IPv6 transition
problem then you should probably leave this list and work on operational
issues of interworking between two Internets, one IPv4 and one IPv6. All
we can do here is to keep a sane, balanced resource allocation system
functioning throughout the IPv4 endgame. That, in itself, does little to
help transition to IPv6.

--Michael Dillon