[ppml] 2005-1:Business Need for PI Assignments

[David Conrad]

Michael,

On Apr 25, 2005, at 8:31 AM, Michael.Dillon at radianz.com wrote:
> IPv6 addresses may
> be finite in number, like water molecules on this planet, but they are
> as numerous as the raindrops which fall on the western rainforests of
> North America.

Under current allocation policies there is a theoretical maximum 
281,474,976,710,656 "allocatable units" and there can a be a 
theoretical maximum of 4,294,967,296 "ISPs", each of which having a 
theoretical maximum of 65,536 allocatable units.

As we have seen with IPv4, 2^32 can be quite limiting if care isn't 
taken.

Of course, there have already been quite a few allocations shorter than 
/32; the shortest in the APNIC and RIPE databases appears to be a /19 
(I find this somewhat ... ironic given the initial allocations made 
when RIPE-NCC was set up was a /19...).

> You are at liberty to run your home network the way that you want to
> but when we make IPv6 addressing policy we have to do it in the context
> of the IPv6 RFCs in which it is mandatory to assign a /48 to any
> stub network unless it needs more addresses. We can't change that.

Can't we?

Rgds,
-drc