Hi Randy, If you check my Sept 4th message to the list, I did mentioned the certainty argument and others, I will attach that message to the end of this one.

Referring the RIR shopping, I do understand the "fear" but I do not believe it is going to happen extensively, and here are the reason:
    - I believe all the RIR staffs do a good job applying their policies and checking the needs of their members.
    - if we do apply conservatives policies to the last allocations, this controls from the staff will be more intense.
    - the size of allocation space that we are proposing (2x/8) is not so "disproportional" in a 2010 scenario, will be less than one year for every RIR, specially with new projects and technologies (fiber in east africa, wimax).
    - Bigger RIR first and companies their-selves at companies latter after the central pool is empty will be able to look at their own pools and look for address economies, because they were allocated much bigger spaces that smaller RIR. So studying each /8 in detail they probably will find unallocated space that will be easier to get for them that going to another RIR. The question is, why not doing so now? because the current policy do not request you to do it, and if no-one else does it, why should I?. That is why I like how our proposal is written as it establishes that the last allocation is the start of a "new phase".



- First of all I believe it is good that we discuss this issue inside the RIR community, specially because if we decide to do something this is the moment. So I believe it is good "institutionally" for the RIR system to be able to discuss this issue. 

- The policy that has been processed is clearly in the arena of what can be implemented easily, with no technical impact on the global routing table. So, what about the option of leaving the allocations as they are?.  It is clear the the actual scenario rewards the utilization of addresses. Is this bad? not necessarily, the central pool will dry at some point.  However, I do see an issue with the uncertainty for the different RIR about the scenario that they will be facing at the time of the last allocation. Different scenarios can be looked if the last allocation is given to "big" or "small" RIR. 

- Also, if any RIR would like to implement conservative policies, that will not happen under the current policy. Conservative policies could be seen as extending the life of v4 or actually encouraging the use of v6 (as David's Conrad slow landing proposal). 

- Another issue is the pressure on the central pool. I believe that David Conrad mentioned several times in different mailing legal issues that IANA  could when facing a last allocation. I do not know if this issue has been studied by IANA (would be good to know) but it is clear that a global policy directing IANA about what to do with that final allocation in terms that we as a global community have consensus is beneficial.

- Now I have to quote (although I do not remember the exact words) a comment given at the mike at LACNIC open policy forum. "we all are in the same lake, a lake that is running out of water, a lake full of sharks but we are all carrying different load in our boats". So we need to look at the allocated address up into this point by the RIR and the legacy allocations.

We know that the smallest RIR are the ones with the lower rate of addresses per user, mainly because they did not get legacy allocations and also because of more extensive use of NAT (I just been in El Salvador and all their university system runs inside a NAT cloud).  But this regions are reaching a big technology breakthrough as WIMAX and 3G internet access will allow broadband access as an alternative to DSL/CABLE or Satellite or just nothing...

So we know that probably ARIN and RIPE will be able to get some legacy allocations back to their pool and will be able to face the beginning of the  transition from the dual stack scenario. But what about APNIC and the smaller RIR?.

Well for LACNIC and AFRINIC clearly it is critical that they are not the first RIRs to run out of addresses as they do not have any negotiation power on a secondary address market nor with hard/soft. manufacturers.


On Wed, 2007-09-12 at 23:35 -1000, Randy Bush wrote:
[ co-chair hat off ]

further, i talked to folk about this and heard two counter arguments.

smaller rirs will get disproportionate space, more than they would soon
consume.  as it will be more difficult for users and isps in those
regions to move to v6, this is probably a good thing.

the smaller rirs will have more space and this will encourage 'rir
shopping', where a greedy company sets up a phony lir in the region just
to get the address space.  i am not sure if this is a realistic fear.
but i do worry it is like saying "the poor person can not defend the
food we put in their bowl.  therefore we should starve them so that a
thief may not prosper."

i would be interested in hearing other pro and con discussion, as i
assume many other people would be.


Roque Gagliano ANTEL - URUGUAY