[Wg-apnic-fees] Re: [sig-policy] The fees and slowness of policy.

[Randy Bush <randy@psg.com>]

[ please excuse occasional humor ]

after some side discussion with my esteemed colleague ming-cheng
liang of twnic, i decided that we needed some actual data on which
to base our discussions.  so i asked apnic staff for some graphs.
i now have them, found them very interesting, and have put them up
for your view.  they are

  <http://rip.psg.com/~randy/061010.apnic/> (click through to zoom)

  o the rate of growth of ip allocations by apnic in the period
    1993-2006.  this is not surprising.  growth has been serious
    and often above linear.

  o the employee count at apnic.  two interesting points here.  in
    the first three years, there was extravagant 50% growth, from
    two employees to three:).  from 98 to 2002, growth looks very
    much like the service growth in allocations, but then flattened
    (dot.bomb?).

so, if income from services kept climbing, and employee growth
flattened, i had to wonder what caused the increasing budget
slippage starting in 2002-2003 to the present.

  o the third picture shows the exchange rate of USD, which is what
    apnic's income is denominated, and AUD in which apnic incurs
    expense.  i think we have a pretty good explanation in this
    simple slide.  the losses are directly explained by the serious
    improvelemt in the australian dollar.

the most amusing solution, and one i vastly prefer, would be for
apnic to move to hawai`i so expenses would be incurred in the same
currency as the income is received.  but somehow, i suspect this
will not be popular with the membership.  but perhaps this would
gain popularity if the members considered that they might suddenly
need to visit apnic headquarters at least once a year. :)

another, boring and much less interesting, approach might be to
build automatic exchange rate adjustment into the fee structure.
while this might work, it would lead to even more unpredictable
fees and with very non-integer constants in the fee calculations.

another approach might be to adjust fees annually to meet exchange
conditions and also take into account other fundamentals such as
expanded or reduced service types (say, for example, we ask for
expensive automated tools to be developed) or volumes (significant
change in allocation service requests).

i am sure there are other approaches.  and these are very much
worth discussing, and i suggest we do so.  but i suspect what is
not really worth much further discussion is why or whether fees
need to go up.

randy