[arin-ppml] Policy Proposal 107: Rework of IPv6 assignment criteria

[William Herrin]

On Thu, Jan 14, 2010 at 6:06 PM, Member Services <info at arin.net> wrote:
> Policy Proposal 107: Rework of IPv6 assignment criteria

As I read it, these are the main things accomplished by proposal 107:

1. Multihomed organizations now qualify for an ARIN /48 based solely
on the fact that they're multihomed. This corrects the serious
technical flaw in current policy where no IPv6 equivalent of NRPM
4.2.3.6 is usable in IPv6 for any practical definition of "usable."

2. Provides explicit address assignments for non-connected networks,
supplementing ULA.

3. Removes the hard dependency on IPv4 policy for determining
qualification for IPv6 end-user assignments by spelling out all the
other reasonable criteria for qualification.

What did I miss?


I offer the following comments:

1. Proposal 106 is superior to and incompatible with proposal 107. I
strongly prefer proposal 106.


2. I'm concerned about assignments to non-connected networks where
qualification is based on the promise that they won't ever connect to
the Internet and therefore won't introduce a route into the IPv6
backbone. If the promise is meant to be kept, I don't think such
assignments should be made from address blocks within 2000::/3.
2000::/3 is intended to be the block used on the public Internet.

Can ARIN readily acquire an address block outside of 2000::/3 for
these assignments? Or perhaps assert a non-binding registry over a
32-bit section of ULA space?

Let me be clear: I do not object to the use of 2000::/3 space for
non-connected networks. I'm only concerned about the non-connectedness
of a network qualifying its registrant for 2000::/3 addresses for
which it would not otherwise qualify. I worry that will either lead to
an end-run around the qualifications analysis for routed space or
result in such a stringent review and high cost as to render the
process useless for the non-connected networks which need addresses.
Either result is a failure.

I agree in principle with a registry for non-connected networks. ULA's
statistical collision avoidance is not as effective as it appears.


3. I observe that advancing proposal 107 in parallel with 106 would
avoid the potentially Faustian bargain of only correcting current IPv6
policy's obvious failings if folks also accept innovations like pools
of fixed-netmask assignments.


If the proposed 6.5.8.3 can be corrected and only if proposal 106
fails to achieve consensus, I will support proposal 107.

Regards,
Bill Herrin



-- 
William D. Herrin ................ herrin at dirtside.com  bill at herrin.us
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
Falls Church, VA 22042-3004