[ppml] Policy Proposal: Decreasing Exponential Rationingof IPv4 IP Addresses

[michael.dillon at bt.com]

> The only thing that looks pretty solid is that 
> there's almost always growth in the number of addresses used 
> up per year. Anything else only persists for a few years and 
> then changes.

Those things that persist for a few years are the trends. And the
changes have been explained due to macro events (CIDR introduction,
telecoms collapse) so even though the trends go through step changes,
they are still trends. One could still do a worst-case scenario based on
the pre telecoms collapse trend and then estimate the probability that
macroeconomic events will lead to that scenario. You then have an
estimated runout date, and a probability that it will occur.

> I don't think my number is very credible either, but I'm far 
> from convinced that the others are better.

This is why I provided URLs to various forecasting methodology sites.
There are some very learned people who have spent their entire life
studying forecasting methodology. If someone were to point out flaws in
Tony's or Geoff's methodology based on the published literature, that
would be valuable. I would also expect Tony and Geoff to fix those
flaws.

In addition, given that there is an accepted methodology for combining
multiple forecasts to improve accuracy and that combining forecasts has
the most success when both the data sources and the methodology is
different, I was hoping that someone like you would take up the
challenge of finding another data source and producing yet another
forecast.

It's easy to wave your hands and be a sceptic but that doesn't lead us
anywhere.

> The theme seems to be either stick close to last year or jump 
> by more than 50%. Not exactly smooth.

What is the frequency of the jumps? What is the average 5-age increase
if you break the year-to-year increases into two classes, small ones,
and bigger jumps.

--Michael Dillon