[arin-ppml] 2008-3 Support

[Milton L Mueller]

> -----Original Message-----
> 
> I don't think that "wanting" an IPv6
> allocation is sufficient. There is an IPv6 addressing model which
> is based on the finite size of the IPv6 address space. It allows
> for nearly limitless growth at several levels of the model, for
> instance a /64 per subnet, and a /48 per site. But at the network
> provider level, the supply of /32s is somewhat more limited. Here
> we want to restrict /32s to those organizations that provide
> interconnect services to many other networks/sites as a major part 
> of their business model. Universities do this, ISPs do this, even 
> the IT department

Sure, a /32 is a pretty big chunk of change and should not be handed out willy-nilly. But if your fee structure is right, those who get more pay more, and the small community orgs wouldn't ask for /32s and incur such fees unless they really needed them. The issue is whether they can get a /48 or /56, just because they want one. 

> > I do think, however, that this experience ought to prompt
> > ARIN to consider more carefully what it means by "need" for
> > addresses in the v6 environment.
> 
> Yes, I agree, and I think that all of us should consider this
> issue, especially in light of the IPv6 addressing framework,
> its current design, and the historical path by which it got
> to that design.

Good to see this. Any more specific ideas? 

--MM