[arin-ppml] Not petitioning proposal 103

William Herrin bill at herrin.us
Fri Jan 1 02:17:24 EST 2010

Happy New Year,

On consideration, I won't petition the AC's abandonment of proposal
103. I accept the AC's conclusion that there is more that can be done
in the process before pushing a formal proposal.

I'd like to thank each of you who committed privately to support such
a petition. Although we could have forced it forward, I think I spy
some potentially more productive avenues for advancement. Rest
assured: I have no intention of dropping the matter.

I enthusiastically support proposal 106. Scott incorporated those
parts of 103's improvements that do not depend on abandoning
needs-based allocation. I think he did a good job and I think the
result would significant;y improve ARIN's IPv6 address management
process. I'll be very disappointed if it is not accepted to the AC's
docket for formal discussion and presentation at the meeting.

Proposal 103 raise another issue that I believe we should address as a
community: should ARIN be the gatekeeper to Internet routing policy?

In most of the time since ARIN's inception, the question has been
moot: network operators depended on ARIN to act as a check on routing
table growth lest the Internet collapse. With IPv4, no credible method
has been proposed for controlling route table growth without RIR
policy enforcement. It's a bona fide "tragedy of the commons" problem.

Proposal 103 offered a credible way to maintain control of IPv6 route
table growth without requiring ARIN to act as the gatekeeper to
Internet routing in North America. Now that we know there's one way,
we may quickly discover that there are more.

So, now that we actually have a choice, should ARIN be set IPv6
Internet routing policy for North America? Or would we all be better
off letting the ISPs individually decide for themselves?

I'd like to see further discussion, particularly about the "how."
Proposal 103 did it with ARIN fees used as a proxy to classify the
network registrant's perception of the value of his system. What other
methods might work?

I'd also like to see all of this presented at this year's meetings.
Judging from the response to proposal 103, the topic has wide enough
interest and support that

My choice comes with two caveats:

1. Although Scott's proposal does not remove ARIN from the position of
setting Internet routing policy for North America, it does incorporate
many other valuable improvements from proposal 103. I believe that
106's adoption would substantially improve ARIN's IPv6 address
allocation policy and practices.

2. I would like to solicit the AC's assistance both in exploring how
to disentangle ARIN from setting IPv6 routing policy and determining
the breadth of support for such a radical departure from ARIN's
current practice.

Since its inception, ARIN has largely set IPv4 routing policy for
North America. It does so by determining (through "need-based"
criteria) who is and is not qualified to obtain DFZ-routeable IP
addresses. ARIN has been thwarted only to a minor extent by the
presence and action of the pre-ARIN "legacy registrants." Such a
top-down and somewhat unintentional routing policy has significant
drawbacks, particularly as it pertains to innovative use by small
organizations. This may not be the most healthy way to build Internet
routing policy.

Proposal 103 was the first time I've seen a credible alternative to
ARIN-set routing policy given serious consideration. The comments
about proposal 103 suggest that there may be wide support for leaving
routing policy to the network operators instead of ARIN.

In addition to discussion here on PPML, I'd like to see ideas for
bottom-up operator-implemented routing policy presented at ARIN's 2009
public meetings and discussed with the wider policy community there
along with any frameworks for actually making it happen. I'd also like
to see that wider community solicited for feedback: was this just a
blip on the PPML or is it something many of them want too?

Given the radical nature of the idea and the detail-oriented nature of
potential implementation frameworks, I believe such a presentation is
beyond the scope of the open policy hour. Short of advancing it as
implemented in a formal policy proposal, how do we get it added to the
meeting agendas?

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

More information about the ARIN-PPML mailing list