[arin-ppml] "Millions of Internet Addresses Are Lying Idle" (slashdot)

[Paul Vixie]

> > no action of that kind can change the inevitability of runout, nor
> > change the short term expedients for all connected networks.  so, why
> > bother?
> 
> I saw a lot of projections about what bigger providers would be forced to
> do in short-term runout, prior to wide availability of IPv6.  Not a
> single person in the room said that multi-level provider-wide NATs would
> make the world a better place.

and yet that is also inevitable.  in the theoretical case you cite where some
currently unused space is put into use (some say by reclaimation, others say
by return, others by a "market") it changes only the date at which runout will
occur, not the inevitability of such a runout.  sooner or later we get runout
and it will bring with it deaggregation, double/triple NAT, and panicked IPv6.

> The amount of effort involved in reclaiming completely unused blocks is
> <= one full time person for a year.  That's a small cost for an
> extension.

i fear that at this point the stage is set by monetary expecations, and those
address blocks which could if recirculated set back the runout date by months
or at most a year, will be closely held by those hoping that a market appears.
to expect that one full time person working less than one year would be able
to counter and/or reset that expectation seems incredibly unrealistic.

this is why to me the most salient of underexposed positions among those with
runout proposals is: do they consider IPv6 inevitable, or do they think that
some combination of deaggregation plus NAT plus their proposal could make IPv4
last forever or at least last until something other than IPv6 can develope?
i really think the community deserves to know/evaluate/choose *that* agenda.