[arin-ppml] Draft Policy ARIN-2024-8: Restrict the Largest Initial IPv6 Allocation to /20

[John Santos]

On 6/28/2024 12:17 PM, David Farmer via ARIN-PPML wrote:
> 
> 
> On Fri, Jun 28, 2024 at 10:01 AM William Herrin <bill at herrin.us 
> <mailto:bill at herrin.us>> wrote:
> 
>     On Thu, Jun 27, 2024 at 4:17 PM David Farmer <farmer at umn.edu
>     <mailto:farmer at umn.edu>> wrote:
>      > On Thu, Jun 27, 2024 at 5:04 PM William Herrin <bill at herrin.us
>     <mailto:bill at herrin.us>> wrote:
>      >> we know a /16 has been allocated. We can't know how they justified it
>      >> because that information is private. Can you produce a -notional-
>      >> justification for a /16 that we all agree is -reasonable-?
>      >
>      > The current policy has been in effect since ARIN-2011-3 was implemented[..]
> 
>     Yes, yes, only one registrant has thus far had the chutzpah to seek
>     and acquire a /16. I have already acknowledged the truth of that
>     claim; you need not continue repeating it.
> 
>     Perhaps you could stop deflecting the question I asked you in return:
>     Do you, David Farmer, believe there exists a justification for an
>     *initial* allocation of a /16 of IPv6 addresses which would withstand
>     public scrutiny? An allocation to an organization which has never
>     before held ARIN IPv6 addresses. If you do, would you care to offer us
>     such a hypothetical to examine?
> 
> I considered this question back in 2011 when the question of /16 or /20 came up 
> in the discussion of ARIN-2011-3. I concluded it was possible to justify a /16. 
> Let me put the question slightly differently: Is it possible to justify more 
> than a /20? There were already /19s allocated by other RIRs, so I concluded that 
> it is possible to justify more than a /20. I also believe nibble alignment is 
> important, so I support /16 as the maximum allocation. Nevertheless, such /16 
> allocation should be rare; one in a decade aligns with that belief.
> 
> However, those who think it is impossible to justify a /16 for an initial 
> allocation should support this policy.
> 
> Thanks.
> 

This is a strawman.  I don't think anyone is claiming it is impossible to 
justify a /16 for an initial allocation.  What people are saying is that they 
don't know of any justification, and without some rational, practical reason why 
someone would need it, we should not grant blanket permission for such a large 
allocation.

Not knowing if something is true is entirely different from knowing it is not true.

The prudent course of action is to set a reasonable limit and if someone has a 
reasonable need to exceed that limit, let them justify that need.  I don't think 
we need to know who they are or the details of their business plan (though if 
they are allocated that space, their identity becomes public record in most or 
all cases), we just want to understand the technical reasons.  Maybe there is an 
alternate method that doesn't require so many bits.  Maybe IPv6 128-bits are not 
enough and we need to invent IPv7 with 4096-bit addresses!

I think it is perfectly rational for people who have gone through IPv4 runout 
and were promised that that will never happen with IPv6 to have some trepidation 
when (if there is exponential growth* of whatever the /16 holder is doing) it 
looks like runout could happen much faster than anyone ever imagined.

[*] exponential growth is the natural result of any successful enterprise until 
it hits a wall and crashes and burns.  I don't want that wall to be IPv6 runout 
because the collateral damage would be devastating.

-- 
John Santos
Evans Griffiths & Hart, Inc.
781-861-0670 ext 539