[arin-ppml] Q1 - ARIN address transfer policy

[John Curran]

At 12:28 PM -0700 6/21/08, k claffy wrote:
>if we're so great at stewardship, how did we get so stuck so fast?

RFC 1669:  The community opted to not to differentiate IPng (now
IPv6) in terms of desirable end-user functionality, and that means
that its deployment needs to be pushed by ISP's to solve their IPv4
depletion problem instead of deployed through end-user demand.

>and why are we now helping
>OECD call in meatspace regulatory systems to help deploy ipv6
>http://www.oecd.org/dataoecd/7/1/40605942.pdf since our
>well-stacked system is failing to get it deployed?

See above.  There's going to be quite a few parties interested in
how the global ISP community makes this happen over the next
few years, particularly given the economic impact of getting it
wrong.

>what about the miracle that has to happen for the current
>routing and economic architectures to accommodate a couple
>decades years of ipv6+ipv4 growth+splintering?

IPv6 routes are bigger, but there are several orders of magnitude
fewer of them today.  We certainly have the ability through policy
to keep the number of IPv6 routes manageable if the community
desires such.

The IPv4 fragmentation impact is still open to debate, and likely
quite variable depending on the introduction of "open" transfers
and the exact policies adopted (if any).

I'll maintain that today's policies for IPv4 and IPv6 allocation provide
an option for growth without routing collapse, presuming you believe
the equipment community, and we have a coordinated effort to move
public Internet servers to IPv6, and the community maintains its
conservative (in terms of routing impact) stance on IPv4 and IPv6
allocation policies.  While I don't have rigorous analysis for the above
personal statement, VAF's analysis from the IAB RAWS workshop
<http://www.iab.org/about/workshops/routingandaddressing/vaf-iab-raws.pdf>
suggest that staying the course for immediate (5 year) future is
uncomfortable but manageable.  The medium-term outlook may even
require reducing IPv4 routing post-transition into to keep everything
running.  Long-term, we all know we'll need the EID/LOC split being
explored by the IRTF RRG work.

In light our vulnerable situation, it definitely would be nice if there
were "some data to support realistic simulation of routing and
market dynamics" that result from significant departures to our
present IP resource allocation policy.  To date there simply hasn't
been any.  This doesn't mean that we shouldn't necessarily make
such changes, only that it requires an amount of faith given the
stakes involved.

>there is no concerted funded effort to develop a long-term sustainable
>routing and addressing architecture while we argue about how to
>prolong the architecture(s) we all insist are inherently crippled.

Yes, that would be quite valuable.  No, it's not the role of the ARIN,
or the RIR's, or ISP's in general to make this happen.  The IRTF RRG
folks are out there working on this, and this community is waiting
patiently for results.

/John