[ppml] Policy Proposal: 2007-12 IPv4 Countdown Policy Proposal

[Geoff Huston]

At 08:47 AM 22/03/2007, David Conrad wrote:
>Hi,
>
>On Mar 21, 2007, at 1:47 PM, Stephen Sprunk wrote:
> >> People don't like the fact that IPv4 is ending, period.
> > IPv4 isn't ending; it's approaching the inherent limits of growth.
>
>It isn't even that.  The IPv4 _FREE POOL as administered by IANA and
>the RIRs_ is being exhausted.  That's all.
>
>There is lots of unused address space locked away in legacy (and not
>so legacy) allocations.  I imagine that address space is increasingly
>going to come into play as folks find they are not able to obtain
>addresses via "traditional" means.

The data on the IPv4 address pools over time appears to show that 
this might already be happening today.

The total size of the address pool that has been allocated by the 
RIRs but is not visible in the routing table  peaked in July 2005 at 
a pool size that was the equivalent of 49.4 /8's Today that 
"allocated but unadvertised" pool sits at 47.0 /8's. i.e. over the 
past 20 months or so the equivalent of 2.4 /8's, or some 40.2M /32s, 
has come into play in the public Internet as address space advertised 
as reachable in the routing system.  While I have not looked hard at 
the data to determine the precise profile of allocation dates of this 
particular pool of address space, it does appear that most of this 
allocated but unadvertised space that has appeared in the routing 
system over this period was originally allocated in the period 1990 - 
1995, providing a strong hint that its the legacy address space that 
is reappearing in this manner already.

[The reports of IPv4 consumption are at http://ipv4.potaroo.net, and 
the time series of the size of the "allocated but unadvertised" 
address pool is graphed in Figure 30b, and allocation date 
distribution of unadvertised address space is shown in Figure 14.]


Geoff