[ppml] IPv4 wind-down

Tony Hain alh-ietf at tndh.net
Wed Mar 21 13:21:24 EDT 2007

michael.dillon at bt.com wrote:
> ...
> However, it would be good to have a reclamation discussion in some
> forum. There are varying levels of reclamation. IANA/ICANN could
> reclaim
> some space such as class E addresses, some of the loopback block, 

This would be pointless. While I don't want this to devolve into MSFT
bashing, every version of windows from 95 through Vista considers the E
block to be insufficiently defined and they refuse to accept configuration
into that space. While you could allocate from the Class E space, when 9x%
of the end systems refuse to use it there is no point.  

> and
> portions of early network number allocations that are not fully used.

Shortly we will be burning through 2+ /8's per month, so the logistics of
reclaiming space will be challenged to keep up, even if there are blocks
that add up to that much. This statement also assumes that just because part
of a block is currently unused that the assignee will be willing to give up
their current 'holding', just for the right to get in line with everyone
else when they need to grow shortly down the road. 

> The RIRs can reclaim addresses by auditing all allocees and assignees
> to
> see if they still are using those addresses. Individual companies can
> reclaim addresses, perhaps by shuffling them around internally with
> lots
> of internal /32 routes. Or by renumbering to get a better internal
> aggregation scheme in place. 

This statement assumes an organization would undertake substantial cost,
strictly to allow a potential competitor to get space. Reality???

> These possibilities whould all be
> documented, along with estimates of the amount of addresses
> recoverable.
> In some cases we might be able to publish some best practices for
> reclamation and reuse of addresses in individual networks.
> > - Encouraging people to use less IPv4 addresses, including but not
> >   limited to higher fees, required use of NAT,
> ARIN should never promote shifting to IPv4 NAT without promoting a
> shift
> to non-NAT IPv6 at the same time. Yes, there are tradeoffs and the IPv4
> NAT solution may be more expedient, however both options should be
> fairly presented. I don't want to see ARIN pushing NAT.

I agree that ARIN should not be promoting NAT. At the same time, ARIN should
not even be implying that they will be miserly with the remaining IPv4
space. Stewardship does not equate to rationing that causes people to do
unnatural things like NAT in their network deployment. It is long past time
to get over it and realize that further extending the lifetime of IPv4 is a
waste of everyone's time and energy. People will need to learn the realities
of the new version of IP, or find another line of work. 


More information about the ARIN-PPML mailing list