[arin-ppml] Revised -- Policy Proposal 2009-4: IPv4 RecoveryFund

[Owen DeLong]

On Apr 13, 2009, at 1:55 PM, Kevin Kargel wrote:

> My apologies for top loading, but it seemed best in this instance.
>
> In first look at 2009-4 I do not hate it, (which I know will shock  
> many of
> you).  To my mind there are a few requirements for any transfer  
> proposal
> that seem to be met by 2009-4.
>
> 1.  ARIN should do any brokering.
> 2.  Anonymity must be maintained, the receiver should not know which  
> IP's
> they are getting or being considered for ahead of time in order to  
> preclude
> side deals and a grey market. Endpoints must not be allowed to  
> negotiate.
>
> One issue I do have is that the proposal is heavily slanted toward  
> the big
> organizations with section 4.X.4 where it states "Recovered IPv4  
> number
> resources should be broken into smaller blocks only if there are no
> bidders for the larger sized blocks"
> I understand the rationale for this, but it would guarantee the big  
> ISP's
> first crack at any blocks leaving only leftovers for small and  
> middle sized
> operations.  This is not a level playing field.

Realistically, I don't think this is an issue. I think the majority of  
the blocks
likely to be recovered will be /18 and longer, leaving the vast majority
of large ISPs out in the cold. I suspect that there will be more /22s  
than
any other size and more /20s than /18s.

>
> I believe that ARIN will tend not to de-aggregate blocks any more  
> than is
> necessary and we can leave this operational feature to the  
> prerogative of
> staff.  It does not need to be in Policy.
> I agree that the statement " ARIN should take all practical steps to
> aggregate returned address blocks." should remain, but the rest of the
> section can be managed by staff at their discretion.
>
This, however, is  valid in my opinion.

> I would suggest that the data reporting for section 4.X.5 be  
> published when
> or after it is thirty days old.
>
Is there a reason for this? I think that publishing 9/1-9/30 on 10/5  
or so
is probably fine, for example.

> I might add a section where IP seekers who advertise or solicit  
> offers be
> disqualified from the process.  Without this there is the  
> probability of
> seeker publishing a general notice that they would bid a million for  
> a /8,
> allowing the returner to submit a safe million dollar offering.   
> This would
> be just a hidden version of peer-to-peer sales.
>
Not so much.  After all, you know that there may be several potential  
seekers
going after that single /8.  Also, the seeker would still have to meet  
ARIN
requirements for a /8, and, frankly, at that level, all the possible  
players
know all the other possible players anyway.  When you get to the realm
of longer prefixes, the number of players expands rapidly and the  
likelihood
of successful peer-to-peer bargaining becomes virtually zero.

Owen