Sheesh

[David R. Conrad]

[Note cc and reply-to]

Brett,

>Not to sound like Mr. Flemming but one thing strikes me as a serious
>oversight,

I don't think anyone sounds like Mr. Flemming -- he is ... amazing.

>Why isn't something being done to create something more scalable and useable?

The less cynical among us would point to IPv6...

>ARIN is _NOT_ the solution.

Perhaps this might be surprising, but I agree.  I feel the current
registry system is a stopgap until "something better" comes along (as
an aside, I had high hopes for automagic address assignment in v6, but
it would appear people still think having strings of (hex) digits in
configuration files is necessary).  The reason for PAGAN is
specifically to address how Internet address allocation will occur in
the future.

However, we have a "situation".  Internet address allocations are
performed in a certain way now (don't want to get into whether that
way is "correct" or not -- that's a discussion for another thread) and
as a result, there are certain costs associated with performing
allocation and registration functions.  Somehow those costs have to be
met.  Right now, they are met by a cross subsidization from DNS fees,
however for various (IMHO very good) reasons, there is strong pressure
to decouple address and name allocation.  Further, within the
Americas, the address allocation services are provided by a for-profit
organization under a cooperative agreement with the US National
Science Foundation and that agreement is due to terminate "soon".

As it is assumed to be very important that IP address allocations be
both stable and reliable, the allocator for those addresses must also
be stable and reliable.  I believe ARIN is an attempt to address this
requirement in the context of existing constraints by using models
(read: working code) from the European and Asia Pacific regions who
didn't have Uncle Sam to rely on.

>As the internet grow IP space is only going to become
>more and more scarce and then your troubles will be even more.

Or they will get easier as (to badly mangle words of Geoff Huston) the
registries which are distorting the true market for Internet addresses
becomes less and less a factor in Internet address transactions --
what is now the black market could conceivably become the normal
mechanism in which organizations obtain address space.  No more having
to justify every last host address... of course, likely a per address
charge under standard supply/demand economics too...

Just one of many possible scenarios, however the fact remains that you
are dealing with a finite resource in increasingly high demand.  The
"Tragedy of the Commons" applies to the Internet too.

>Now I'm no
>genius, and I can't provide the answers.. what I want to know is why the
>people that can aren't.

Because it is a very hard problem that [IMHO] is made essentially
intractable by political and religious stances.  There is a tremendous
amount of historical baggage associated with IPv4 which can't simply
be undone by wishing it so.

Regards,
-drc