[arin-ppml] A modest proposal for IPv6 address allocations

[Joe Maimon]

William Herrin wrote:
> So here's a crazy plan:
> 
> A. The first IPv6 allocation from ARIN is always a /48. To justify it,
> you need to be multihomed. There are no other qualifications. The /48
> will be allocated from a pool from which only /48's are allocated.
> 
> B. The second IPv6 allocation from ARIN is always a /32. To justify it
> you need to demonstrate that you've efficiently used the /48 for some
> reasonable definition of efficient, that you've implemented SWIP or
> RWHOIS for your downstream assignments and that you will run out of
> space in the /48 within one year. The /32 will be allocated from a
> pool reserved for allocating /32's.
> 
> C. The third IPv6 allocation from ARIN is always a /24. To justify it

Renumbering is something everyone should desire to avoid, regardless of 
how easy it is, and I would oppose any policy promoting that activity, 
so please clarify whether you were intending for renumbering to occur or 
not.

The only sane method for general prefix filtering is per prefix size per 
identifiable pool.

And there should be only a few of these pools.

So please put a nail into the mantra of registries not being involved in 
routing. It becomes rocket science and/or mind numbing effort to the 
point of impossibility to try to prefix filter the internet without some 
sort of clear guidelines and easy to follow heuristics. And they need to 
stop changing with the seasons.

I suppose secured bgp will fix all that. I bet registries will have even 
more to do with routing under that system. Put it to bed. Its good 
philosophy but bad practice.

All effort should be made to limit the prefixes an ASN has to advertise 
for all its needs. That means to me erring on the side of large and 
sparse allocation techniques.


Joe