[ppml] rubber/road

[Stephen Sprunk]

Thus spake "Ted Mittelstaedt" <tedm at ipinc.net>
> I assume you know this depends on the assumption that all
> non-paying IPv4 legacy holders will start paying the RIRs for
> IPv6 allocations in that way we can spread the costs around
> more.

You assume costs will remain at current levels.  However, yes, eventually 
legacy-holders will need to start paying fees if they want to get direct 
IPv6 assignments.  It remains to be seen if they will within a timeframe 
that merits discussion today.

> If that was not the case for some reason, then it would be
> impossible for v6 to stay cheaper than v4, unless that is, the
> RIR's costs for niceties like heat, light, office space, labor, etc.
> were to somehow decrease.
>
> Nobody will pay for v4 once v6 has taken hold, in that case then
> v6 prices must rise unless cost to administrate numbering goes
> down, somehow.

If budget shortfalls are predicted, the Board is required to increase 
revenue (which probably means raising fees, but might mean lowering them) or 
decrease costs, regardless of what the cause is.  That's not necessarily 
linked to IPv6 deployment or IPv4 becoming worthless, though.  There's also 
no link between those things and the relative fees for IPv6 vs IPv4; the 
Board could instead decide to raise IPv4 fees faster than IPv6 until people 
just stop using IPv4 -- there've been a few informal proposals to do exactly 
that to delay (or even reverse) exhaustion.

> Or is there some magic feature about IPv6 I have missed that
> makes it a lot cheaper to administer, once everyone has
> switched over to it?

The "magic" is that we've tried to make IPv6 policy such that the vast 
majority of orgs will never need more than an initial, minimum-sized block 
which is somewhere between "trivial to justify" and "automatic".  Compare to 
IPv4, where many orgs have _hundreds_ of blocks and have to come back for 
more every few months, requiring detailed review by staff.  Simply put, IPv6 
should require fewer man-hours to handle than IPv4 because the request 
volume will be lower and the effort per request will be lower.  Bad news for 
ARIN staffers, I suppose, but good news for fees.

Some fee adjustment may be necessary for ARIN to keep funding fixed-cost 
items like servers for WHOIS and DNS, internal systems and tools, meetings, 
outreach, etc. if the maintenance fees (plus the occasional new assignment 
fee) for IPv6 won't be enough after IPv4 finally dies, but that won't be 
"jacking" with anyone -- it'll be necessary to remain solvent, and it's far 
enough off in any case that it's not productive to try to predict what's 
going to happen or when.

S

Stephen Sprunk      "Those people who think they know everything
CCIE #3723         are a great annoyance to those of us who do."
K5SSS                                             --Isaac Asimov