[arin-ppml] RIPE/ITU

McTim dogwallah at gmail.com
Fri Feb 26 17:50:37 EST 2010


Hi Milton,

On Sat, Feb 27, 2010 at 1:15 AM, Milton L Mueller <mueller at syr.edu> wrote:
>
> > -----Original Message-----
> > You might want to look into APNIC's NIR model which is exactly
> > that, and then make a formal proposal.
>
> My understanding is that the people at APNIC are not exactly delighted with the NIR model;

My understanding is that the people who make up the APNIC community
are not exactly delighted with the NIR model, and the model is being
phased out as not viable.

>
> it was a concession to some of the same political pressures that are now leading to the call for CIRs. Indeed, it is a bit inconsistent for this community to argue against the Ramadass proposal for CIRs and then propose NIRs instead.

agreed, but the community hasn't endorsed this idea AFAICS.

>
> I prefer McTim's approach of supporting the principle that addresses should be allocated to users (ASs) and not to governments or other entities who want to interpose themselves as intermediaries.
>
> If it comes down to a choice between NIRs and CIRs, then the only relevant difference is whether the RIRs retain a monopoly on higher-level allocations or not. I think that's a weak position to be in. Here we need to frankly realize that the issues are fundamentally political.
>
> Note that the ITU proposal for CIRs does not propose to make them exclusive, but rather proposes that an ITU-mediated CIR be an additional option. If one supports competing ISPs, why not alternative address registries?

RFC2050

"Service areas will be of continental dimensions."

In addition there are various other reasons outlined in the Circleid
post I sent earlier.

>
> One cannot argue against this option on the grounds that we don't have enough ipv6 addresses to make it viable; we do. One cannot argue against it on the grounds that it messes up the efficiency of routing,

You can actually.

>because new, RIR-sanctioned NIRs or new RIRs carved out of existing ones would have basically the same effects on routing.

perhaps, or perhaps not, depending on the policies they adopt.

>
> The only way to viably challenge this proposal is to face squarely the political issues and question whether states who gain control of ip addresses will truly allow competitive alternatives, or instead try to leverage their control of the resource to gain more control over internet supply and usage, or to favor incumbent, national champion network operators.


or to boost government revenue, or use IP address revenue to feather
individual nests, or not develop policies in an open, transparent
manner, etc, etc.


>
> One can also try to question the need for this proposal (solution in search of a problem). There is merit in this argument, but the issue is not quite as simple as it may seem. Developing countries aware of the landrush for ipv4 addresses that took place in the early stages of the internet's development have a legitimate reason to worry about whether history will repeat itself. An inherent feature of needs-based allocations is that developed economies will be able to show "need" well before undeveloped ones. The ITU is basically arguing for a system of reservations that guarantees each nation-state a substantial chunk of address resources just in case that happens. Naturally enough, given its basis in an intergovernmental organization, the ITU considers nation-states the appropriate stewards for these reservations.


Speaking of "reservations" I am under the impression that it is the
IETF that instructs the IANA to reserve IP blocks.  If this is
correct, then the ITU and the RIR policy lists are not the correct
fora to discuss this proposal.  The ITU should write a draft RFC and
submit it to the IETF.  Perhaps I am mistaken.

>
> If setting aside address block reservations were the ONLY price we had to pay to assuage these political concerns, I would say, "do it." But other prices could be extracted - as I explained above, linking ip address allocations to the nation state system so closely could have serious political and regulatory repercussions. That price is not worth paying.


I hope you plan to say this to the ITU in March!


--
Cheers,

McTim
"A name indicates what we seek. An address indicates where it is. A
route indicates how we get there."  Jon Postel



More information about the ARIN-PPML mailing list