[arin-ppml] Bootstrapping new entrants after IPv4 exhaustion

[Scott Leibrand]

During the discussion in Phoenix of Draft Policy 2013-7 (which I've since
updated and will be sending back out to PPML shortly), and in other
discussions before and since, it has become apparent that small networks
may not qualify for transfers and be unable to get space from their
upstreams after RIR and ISP IPv4 free pools run out.

In order to address this issue, a few different ideas have come up, so I
wanted to bring some of them up to the community for discussion and see
which possible solutions might have community support.

Here are a couple of the ideas that've come up so far:


1) For smaller allocations than ARIN’s minimum, orgs “should request space
from their upstream provider _*or another LIR*_” (add underlined text to NRPM
4.2.1.5 <https://www.arin.net/policy/nrpm.html#four215>).

This would clarify that it's fine for organizations to get space reassigned
to them by any other LIR if their upstream ISPs are no longer able to
provide them the space they need.


2) Lower the minimum allocation sizes to /22 single-homed and /24
multihomed for both ISPs and end-users.  This would mean updating NRPM
4.2.2<https://www.arin.net/policy/nrpm.html#four22> and
4.3.2 <https://www.arin.net/policy/nrpm.html#four32> (and would allow
removal of NRPM 4.9 <https://www.arin.net/policy/nrpm.html#four9> as
redundant.)

Before the implementation of CIDR, many /24 allocations were made to
organizations that are no longer using them.  Current ARIN transfer policy
<https://www.arin.net/policy/nrpm.html#eight3>states that the minimum
transfer size is a /24, but it's not clear in policy whether an
single-homed organization that needs a small block (/24 to /21) would
actually qualify to receive such a block via transfer.  (Perhaps staff
input here would be useful.)  In any event, reducing the minimum allocation
sizes would allow organizations of all types to receive the size of address
block they actually need, either via transfer or from ARIN's inventory of
returned space.

Thoughts?  Do you support either or both of these ideas?  Would one or both
of them be worth submitting as a policy proposal?

Thanks,
Scott
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20131121/9375a1c3/attachment.htm>