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

[Joe Maimon]

michael.dillon at bt.com wrote:
>> 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.  
> 
> No, no, no, no, no!

Yes.

> It is fundamentally
> different from IPv6 and requires a change from IPv4 
> thinking, in order to fully understand it.
> 

No, its just a larger binary number.

More bits, less magic.

Every time the paradigm police come around, it gets smaller.

The part that appears to get people uneasy is that the orders of 
magnitude larger than ipv4 ipv6 is supposed to be keeps shrinking. 
Nobody is concerned that we will face exhaustion in our lifetime. Its 
the grandkids lifetime where the possibility now seems feasible.

Are you advocating for a system of addressing that you can believe will 
be measured in decades or centuries?

2000/3 has 500m /32 (32x the number of ipv4 /24)

Real utilization hasnt started yet and 72779 are allocated.

%0.014

http://portalipv6.lacnic.net/en/ipv6/statistics

Geoff Huston has a couple of excellent articles regarding this, we seem 
to keep revolving around the same thought track.

http://www.ripe.net/ripe/meetings/ripe-50/presentations/ripe50-plenary-wed-ipv6-roundtable-report.pdf

His number was 60 years.

Joe