[arin-ppml] Fairness of banning IPv4 allocationsto somecategoryof organization
David Farmer
farmer at umn.edu
Mon Oct 12 19:20:30 EDT 2009
On 12 Oct 2009 Kevin Kargel wrote:
> Do I remember correctly that there was a discussion about implementing
> a general cap on allocation size that would cover this issue without
> all the politics? If we really want to do this perhaps the
> diminishing cap would solve the problem.
There have been several discussion on PPML and within the AC about
allocation limits, some say they are non-discriminatory others say that they
discriminate against the large users of IP space, especially large ISPs.
At ARIN XXIII in San Antonio, draft policy 2009-2: Depleted IPv4 reserves
was discussed; basically once ARIN dropped below a /9, it limited all
allocations to /20, within a 6 month period. Many felt this rationed IPv4
address space to much and wasn't fair to larger and medium sized ISPs.
https://www.arin.net/policy/proposals/2009_2.html
In response Leo and I provided two IPv4 run-out proposals which have been
merged into a single draft policy 2009-8: Equitable IPv4 Run-Out. Which will
be discussed next week at ARIN XXIV in Dearborn.
https://www.arin.net/policy/proposals/2009_8.html
This draft policy proposes to reduce the supply of IPv4 addresses that can
be requested from a 12 month supply to a 6 month supply and then to a 3
month supply as the IANA free pool dwindles down. Then when ARIN
receives it last /8, set a maximum allocation size of (1/4) one quarter of
ARIN's available IPv4 resources within a three month period.
The intent is to reduce and limit the likely competitive issues created as IPv4
runs-out. If some set of ISPs get a one year supply of addresses right before
run-out and another set of ISPs do not, then this will create competitive
issues within the ISP market. By gradually reducing the length of supply that
can be requested, these competitive issues can be contained to
approximately a quarter or two instead of as much as a year or two that is
likely under current policy.
Further by putting in place a maximum allocation of (1/4) one quarter of the
available resources once ARIN receives its last /8, ensures that no one
request can consume all of ARIN's resources.
It is my belief that this policy is still needs based, it simply recognizes that the
not everyones needs are going to be able to be meet and that some
measure of equity is necessary in addition to a simple needs basis, as the
IPv4 resources are depleted.
While I believe that the Smart Gird folks should be encouraged to use IPv6,
and it should be made plain to them that only IPv6 has the necessary
resources to really meet their long-term needs. I believe they should have
equal access to the remaining IPv4 resources under the same terms as
everyone else. They should have the same limits on the supply they can
request. Therefore, they should only be able to get what they can deploy in
the given time frames and given that they would likely need to go touch each
meter to do this, I think there is a reasonable self-limiting factor on the
amount of address space they can consume before IPv4 is depleted.
I think it is really dangerous to start talking about who is worthy or not worthy
of receiving the remaining IPv4 address space. You could argue that only
new users should get it, you could argue that only current users should get it,
you could argue that only the big ISPs should get it, you could argue that
only the small ISPs should get it, you could argue that only end-user
organizations should get it, and you could argue that only those that have
deployed IPv6 should get it. If we do any of those, I believe all we will do is
waste effort on IPv4 that should be spent on deploying IPv6.
Please understand 2009-8 is not intended to do anything to control the
overall rate at which IPv4 is depleted, its only goals are to reduce and limit
the likely competitive issues as we approach IPv4 depletion and to prevent a
single request from exhausting ARIN's IPv4 resources once ARIN has
received its last /8.
Why is this important? If some believe there is even a medium-term
competitive advantage with IPv4 and to not adopt IPv6, they my delay IPv6
adoption. Also, if there were a nasty surprise and ARIN all of a sudden runs-
out of IPv4 because of a final large request takes everything that is left.
Then effort will be spent arguing and making recriminations about how IPv4
depleted, and not spent on adopting IPv6. So, it is important to maintain
some semblance of order as IPv4 depletes, to reduce distractions and allow
people to focus on IPv6 adoption.
We need to focus on IPv6 adoption, and only make those changes to IPv4
policy that helps us stay focused on IPv6 adoption. Continuing a debate
about which uses justify the limited IPv4 that is left is distracting, and will
likely only fracture our community when we need to be focus on IPv6
adoption.
===============================================
David Farmer Email:farmer at umn.edu
Office of Information Technology
Networking & Telecomunication Services
University of Minnesota Phone: 612-626-0815
2218 University Ave SE Cell: 612-812-9952
Minneapolis, MN 55414-3029 FAX: 612-626-1818
===============================================
More information about the ARIN-PPML
mailing list