[arin-ppml] ARIN-2011-5: Shared Transition Space for IPv4 Address Extension - Last Call

[Benson Schliesser]

> The ARIN Advisory Council (AC) met on 13 April 2011 and decided to send the following draft policy to last call:
> 
>   ARIN-2011-5: Shared Transition Space for IPv4 Address Extension

I support this draft policy.  In an ideal world, this block would have come from IANA - but given our current situation (IANA post-exhaustion) I think it is reasonable for ARIN to take action.

On Apr 19, 2011, at 3:14 PM, George, Wes E IV [NTK] wrote:

> First, the current text does not reflect the valid concerns raised in discussion in PR about a lack of enforcement.
...
> Supporters have been quick to say that this space is being reserved for a narrow application and not a general application like an
> extension of RFC1918, and so if this is the case, then we should be providing clear guidance to staff on this matter. 

My view is that "enforcement" isn't necessary, and indeed may not be possible.  There are technical requirements for globally unique space, including scenarios involving NAT, and we shouldn't ask ARIN staff to pick and choose what's eligible for this Transition Space block.

But the NRPM, upon adopting this policy, should definitely include a statement that it is not to be used like RFC 1918 space.  In other words, it is meant for service providers to deploy on internal infrastructure etc - end sites should not use this block, or they can expect to have problems.  Further, service providers that in-turn have upstream service providers should exercise caution when using this block.

> Lastly, while it didn't come up in PR, we had also discussed on PPML the idea of a different sort of sunset clause, where unless X
> SPs came forward and notified ARIN that they were using the space, it would be returned to the free pool instead of being stranded
> and unused. Because there are many of us that remain unconvinced that this will actually reduce the burn rate of IPv4 space, this
> would be a show of good faith on the part of the beneficiaries of this policy that wouldn't fundamentally alter its intent. It
> does't have to be public, but at least it would provide a way to ensure that we're not wasting space on this application. 

On one hand, I'm with you about not wasting space.  But on the other hand, I don't think that: a) we will care about getting this block back, at the future point in time when it is no longer useful; and b) we can be certain that this block is not being used, after it's been set aside.

Personally, I think we should just consider this block a "gift" to the Internet.  We could perhaps suggest that other RIRs, and maybe the IETF, include this block in their documentation - so that it's used by everybody, for the maximum global benefit, etc.

Cheers,
-Benson