[arin-ppml] ARIN-prop-154 Shared Space for IPv4 Address Extension (w/IETF considerations)

[William Herrin]

On Sun, Jul 3, 2011 at 10:57 PM, John Curran <jcurran at arin.net> wrote:
> On Jul 4, 2011, at 12:26 AM, William Herrin wrote:
>> Would anyone object if I replaced this paragraph with:
>>
>> "The IETF is asked to create an RFC governing the use of the allocated
>> /10 for the purpose described. ARIN shall make resources, such as
>> manpower, available to the IETF as necessary to facilitate such
>> activity."
>>
>> I think the original version reads a little too much like explaining
>> to the IETF, step by step, how to do their job. John Curran has called
>> me out on that, and while I disagree with his general position I do
>> think the text can be worded so that it is less likely to be
>> misinterpreted.
>
>  RFC2860 does not indicate that the IETF will create RFCs that "govern the
>  use of the address space"; it states that the assignment of specialised
>  address blocks shall be done by the IANA, only as directed by the criteria
>  and procedures specified in RFCs. The IAB has already indicated that the
>  allocation would be in conflict with these provisions in RFC2860 if done
>  prior to the establishment of consensus within the IETF in this matter.

John,

Yes, I understand your interpretation of the matter. You recommend we
punt back to the IETF and let them run with the ball for a while. And
if someone else would like to write a proposal that ARIN play lap dog
to the IETF on prop 2011-5, I'm sure I can find it within myself not
to invent process grounds for objecting.

In the mean time, you have a proposal headed for general consensus,
advanced primarily by the very folks who vote for six of the seven
individuals who sit on the ARIN board. The proposal, in essence, is
that ARIN act now and encourage the IETF to catch up when it can.
Where the specific word choice sidetracks us, I'd like to fix it.

So I ask you: does this replacement resolve your concern with the
prior text that it could be viewed as a procedure instead of a policy?
I recognize that you have other concerns, but does it resolve that
one?


On a personal note, I appreciate your engagement on this issue. While
I disagree with some of your conclusions, I have found your comments
to be valuable and insightful.


Not for nothing, but RFC 2860 can reasonably be read to prohibit the
IANA from assigning address blocks to organizations who will employ
them as anycast addresses. You know, folks like Network Solutions and
Neustar. Unlike the shared address space in 2011-5, 2860 explicitly
calls out anycast addresses (section 4.3 paragraph 2(b)). Not only
that, it prohibits IANA-assigned addresses from being used for
experimental protocols that consume IP addresses. 2860 calls
experimental protocol addresses out as well. The IETF makes those
assignments by RFC.

Yet under current ARIN policy, anycast is just another multihomed use
and the addresses Facebook uses with the experimental LISP protocol
are not subject to revocation. Where is the outrage?

You have 2860 wrong. It doesn't prevent ARIN from allocating addresses
for use on the public Internet. It only enjoins ARIN from acting as
the technical standards body that precisely defines the allocation's
use.

We can dance the language a little bit more. Make it an allocation
which will be made in whole to multiple organizations instead of just
one so that in a strict technical reading it is virtually the same as
every other allocation we make. After all, we don't guarantee
routability of any assignment and 2860 doesn't actually require one
registrant to an address block. Global uniqueness is our policy and
while there are RFCs, uniqueness is not part of the IANA MOU.

But wouldn't you rather assume that the folks in the IETF are not
a-holes and instead of trying to foil us will build the shared space
we ask for? I've participated in the IETF and I'd like to think my gut
reaction would be to think that if ARIN is willing to give up the
addresses it must be important enough to do. And wouldn't you rather
take ARIN constituents at their word that they need something usable
sooner than the IETF can do a comprehensive and correct job of
standardizing it?

I would.

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