[ppml] Policy Proposal 2007-21: Last Call

[Owen DeLong]

On Apr 14, 2008, at 8:20 PM, Dale W. Carder wrote:

>
> On Apr 14, 2008, at 5:09 PM, Owen DeLong wrote:
>>> Why should historical IPv4 assignments hold up deployment of IPv6?
>>>
>> People who have them and have the ability to independently multihome
>> with their IPv4 assignments, but, cannot get IPv6 assignments on the
>> same terms are unlikely to be willing to implement IPv6 on their
>> networks as a result. This is a not insignificant portion of networks
>> in North America.
>
> Agreed, but if you want "the same terms", shouldn't
> we take out the requirement for a legacy RSA to be in
> place?
>
There are limits to the extent to which the community is willing
to continue the legacy free-for-all.  Not having an RSA is outside
of those limits.  The IPv6 will have to be covered by an RSA
anyway, so, covering the legacy space with a legacy RSA
doesn't seem a particularly high barrier.

>> For example, a site with
>> 150 hosts in 16 class C network ranges would be expected
>> to pick one class C to keep, renumber the hosts they had to
>> into that range, and, return the others in order to be able to
>> demonstrate "efficient utilization".
>
> I would think that it is a waste of time for anyone
> involved to do any sort of minor IPv4 reclamation and
> further inhibit the world's v6 rollout because of it.
>
There are two ways to approach this.  One is to think (possibly
erroneously) that reclamation somehow delays IPv6 deployment.
The other is to recognize that it won't actually, but, reclaimed
space may be useful in easing the transition, and, eliminating
waste of space in IPv4 in favor of justified need based allocation
is probably a good thing regardless.

Obviously, I have my bias as to which is the correct answer of
the two.


> Maybe I'm thinking too much along the lines of "Have an
> AS?  Here's your /48".
>
Yep.  Not actually a good idea.

Owen