One More Time

[Jim Fleming <JimFleming@doorstep.unety.net>]

On Monday, April 13, 1998 2:41 PM, Avi Freedman[SMTP:freedman@netaxs.com] wrote:
@> > I claim that it is hard to verify.  The work is trivial to appear to be
@> > an ISP, for a huge gain.
@> 
@> What gain?  Address space is not owned, right?  It has no value, right?
@
@Karl, whether it's owned or not, if we run out we could be in deep doodoo.
@

Avi...let me try one more time...here is a small exercise...

Let's say that we have an 8 bit address space with only 256
addresses. Let's say that we decide never to give out a block
smaller than 8 continguous addresses. This would result in
a maximum of 32 blocks. Now, let's say that we give these
32 blocks out to the following 7 companies:

A - 8 blocks
B - 4 blocks
C - 2 blocks
D - 2 blocks
E - 8 blocks
F - 4 blocks
G - 4 blocks

OK...the shelf is empty...except in five minutes we discover
that company D is going out of business and their block is
turned in by companies C and E who watch as part of their
"neighbor net" duties. Those 2 blocks from D can be given to
one new arriving company. The new company may only need
one of the 2 blocks so we may end up with 8 companies in
the list above, or a block on the "shelf" for a short period of
time.

Someone could probably build a simulation of this sort of
checking in and out system. The point I am trying to make is
that there should not be any fear that IP space will not be
there because there is always something changing. In a
large address space, there will always be blocks in one of
four states.

1. In-use
2. Being Turned In (This is where we should be looking)
3. Being Checked Out
4. On the Shelf (This can approach zero and #2 can still feed #3)

What good does it do to have little or no time or money spent
on #2 above and have the shelf full while everyone argues that
#3 should not occur because they do not want people to use
the blocks ?

-
Jim Fleming
Unir Corporation
IBC, Tortola, BVI