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

[Milton L Mueller]

Ah I see. The typical residences will need 4,722,366,482,869,640,000,000 addresses, whereas your typical ASN is more likely to need 1,208,925,819,614,620,000,000,000 right off the bat.  

> -----Original Message-----
> From: Scott Leibrand [mailto:scottleibrand at gmail.com]
> Sent: Monday, June 01, 2009 2:41 PM
> To: Milton L Mueller
> Cc: William Herrin; arin-ppml at arin.net
> Subject: Re: [arin-ppml] A modest proposal for IPv6 address allocations
> 
> FWIW, the IANA only got 1/8 of the IPv6 space. The IETF left the rest
> to future generations.
> 
> /56 makes a lot of sense for residential allocations, but not for
> assignments from ARIN. I think /48+ for end users with ASNs, and /32+
> for LIRs (anyone who reassigns /48s) is still the right way to go. IMO
> we just need to simplify our definitions a bit.
> 
> -Scott
> 
> On May 30, 2009, at 4:48 PM, Milton L Mueller <mueller at syr.edu> wrote:
> 
> >
> >> -----Original Message-----
> >> Not sure about all the details, but I like the fact
> >> that we'd be able to do away with the ISP/end-user distinction,
> >> make it
> >> easy to get a /48, and provide a simple growth path for the most
> >> common
> >> cases...
> >>
> >> -Scott
> >
> > Ditto.
> >
> > But, let me express (uncharacteristically) some concern about overly
> > liberal initial allocations. (e.g., why not a /56?) From the
> > standpoint of developing countries, there is some legitimate concern
> > about reproducing the land rush phase of IPv4 address allocations
> > (oops, there goes 1/3 of the space....)