Encouraging return of legacy space WAS Re: [ppml] ARIN Policy Proposal 2002-9

[J. Scott Marcus]

Quite a few of the responses to my last post dealt with the interaction 
with IPv6, which indeed is what I was referring to obliquely when I asked 
whether deferring the exhaustion of IPv4 space was necessarily a good thing.

I think that it IS a good thing.  I continue to feel that the cost of a 
PILOT PROGRAM to try to voluntarily reclaim IPv4 space is low, and is 
warranted.  Right now, we have no idea what the community response would 
be.  I think that it is worth a try, independent of our assessment of the 
probability of IPv6 deployment.

For reasons of fairness, it is probably not possible to do a statistical 
sample; rather, we would probably have to poll a CLASS of legacy address 
holders, I think, in order to determine the feasibility of recovering 
space.  That is what motivated my proposal to contact the ~50 holders of /8's.

I suspect we would find that a few are heavily used, some only moderately 
used, and probably a few that have little or no usage.  For organizations 
who are making only limited use, and are willing to migrate, we could 
consider cutting them a "deal", in order to make the migration less 
onerous.  We might offer to assign a replacement block, and to waive or 
reduce fees on the block for some number of years.  (No loss of revenue to 
ARIN - they are paying nothing today, anyway.)  Perhaps we would need to do 
something to increase their confidence that they would not have problems if 
they later found that they needed more space...  All of this needs thought.

I think that we would also want to publicly recognize their 
public-spiritedness, assuming they want the recognition.  Maybe put them on 
a list of good netizens on the ARIN website, put out a press release, give 
them a trophy or a medal...  You get the picture.  It's good for them, and 
it also encourages others to do the right thing.

These issues are a bit similar to those that the FCC encounters when we 
move folks from one chunk of spectrum to another - we have to give them 
time to migrate, and we may need to find ways to lessen the pain.  But we 
hopefully provide overall benefits by moving a limited resource to higher 
value use.  Analogously, reclaiming underutilized IPv4 space could 
ultimately benefit the whole Internet community.

I expected that somebody would argue that IPv4 reclamation would reduce 
incentives to migrate to IPv6.  To me, that's a real concern, and one of 
the largest downside risks.  I think that there are simple ways to address 
this - for example, we might agree in advance that any reclaimed address 
blocks would be set aside as a reserve for a rainy day, or perhaps as a 
reserve to facilitate an eventual IPng migration.  Who would hold that 
reserve (ARIN, the RIRs collectively, ...) is TBD.

My two cents,
- Scott