[AusNOG] IPv4

[Mark Newton]

On 06/03/2013, at 12:44 PM, Noel Butler <noel.butler at ausics.net> wrote:

> On Tue, 2013-03-05 at 09:44 +1100, Mark Andrews wrote:
>> 
>> In message <1362435579.7275.5.camel at tardis>, Noel Butler writes:
>> > 
>> > On Mon, 2013-03-04 at 21:28 +0000, Bevan Slattery wrote:
>> > All the ipv6 fanbois here today going  we will never run out blah blah
>> > blah should remember one thing (especially dishing out /64's to end
>> > users) ...  I'm sure that's exactly the same train of thought the ipv4
>> > guys had twenty years ago.
>> 
>> And you would be wrong.  The designers of IPv4 knew that 32 was not
> 
> 
> I guess we will see,  in twenty years from now


20 years is 7300 days.

The IPv4 Internet has 2^32 addresses.

Lets say we were going to assign a /64 IPv6 subnet for every IPv4 address.  And furthermore,
let's say we were going to do it again every single day for 20 years.

At the end, we'd have assigned 31,353,261,260,800 IPv6 /64 subnets.

The 128 bit address space provides for 18,446,744,073,709,551,616 IPv6 /64 subnets.
At 2^32 /64 subnets per day, that's enough for 4,294,967,296 days, equivalent to
nearly 12 million years.

I think we'll have enough.  We'll have enough even if we egregiously waste them.  And
IANA has only released 25% of the total address space thus far, so if it turns out that
we burn through that in less than a century we'll be able to pick a different, slower
allocation strategy for the remaining 75%.

  - mark