[ppml] Free Market

Paul Vixie paul at vix.com
Sat Aug 25 12:34:49 EDT 2007

> I have a proposal. In an earlier e-mail someone made mention of the word 
> 'oil'. IPv4 addresses are like oil, in that it is a finite resource, and 
> one day... we will run out. The way our economic markets handle this is 
> simple. Increase the price of oil. Allow economic demand to set the 
> price, and as the price of oil increases, other, often environmentally 
> friendlier technologies take hold.

i think we'll be choking on the fumes of the oil we've already burned long
before we can use up the earth's supply, which ought to counterbalance the
price index.  there's a fire burning in greece right now that they just
can't seem to put out, and the iberian peninsula is in a sustained drought,
and arctic ice is melting, and so on.  burning the last drop of oil, at any
price, seems unlikely at this stage.  i mention this only because it's an
absolutely perfect analogy to the real situation with IPv4.

> What if we increased the price of new ipv4 allocations, across the board 
> - and altered the current pricing scheme, such that those with legacy 
> /8s pay what they are truly worth. This would be an incentive for those 
> who have large volumes of unused space just sitting there to get rid of 
> it, but also allow growth and use for those who are willing to pay for 
> it. IPv6 prices, of course, remain forever at rock bottom.

the limitation isn't on address space, it's on routing table size.  using
the last ipv4 address will require that advertisements be much smaller than
they are now, which would mean a routing table much larger than we have now.
so, just as with oil the long term problem is carbon not supply, with IP
the long term problem is routing not supply.

> If you inserted economics into it - you might see some large blocks being
> returned. I know of one ISP here in Lexington that has 130,000 legacy IPv4
> addresses. They use... maybe 2000. In fact, if Arin were to turn down a
> request for me - they'd be my first stop. I'd buy a t-1, with a /19
> attached. I can't use the T-1 of bandwidth - but I'd certainly advertize
> the space out my Gig-E or OC12 links. That is an example of what some of
> us have to do to get additional IPv4 space in a hurry.

how many /16's can be carved up into /19's before the global routing table
exceeds the size of routers that other networks can afford?  and if we get
to that point, will it make sense to add another /19, no matter where you
can get it from or at what price, given that it won't be globally reachable?

> I know starting up, we had to get our address space from upstreams. When 
> we'd cancel an upstream... .renumbering time! Now that we have ARIN 
> space, we don't have to do that - but if ARIN can't assign it - I'll 
> certainly 'buy' it from someone else using the method detailed above. 
> Paying monthly for "bandwidth" but actually for IP space.

so, provider-assigned (PA) has a provably bad short term economic model (it
functions as a price lock due to its renumbering penalty), whereas provider-
independent (PI) has a provably bad long term economic model (it blows out
the routing table).  which bad economic models would ARIN's community like
to pursue?

> So, one way or another when IPv4 becomes scares, restricted or 
> unavailable, the market will handle it.

by moving the problem from "us" to "our grandchildren".  just like with oil.

> Why not shift the burden now onto those who hold all those massive
> allocations, and use ARIN to regulate the price of IP space.

that's actually an unrelated topic.  nobody anywhere has the authority to
impose revocation terms or prices on pre-RIR allocations, and that's where
the majority of IPv4 space still is.

More information about the ARIN-PPML mailing list