[arin-ppml] Queue depth report?

[Geoff Huston]

On 1 Oct 2014, at 1:23 am, John Curran <jcurran at arin.net> wrote:

> On Sep 30, 2014, at 10:45 AM, Matthew Kaufman <matthew at matthew.at> wrote:
> 
>> I've been watching Geoff's ARIN runout prediction slip out into the future, and reading here about increased review leading to slower processing... Is runout being delayed just because legitimate applications for space are queuing up?
>> 
>> Can we get a report on historical and current application queue depth from staff?
>  If you need anything additional, please let me know... (Also, I've cc'd 
>  the esteemed Geoff Huston on this email; it's only fair in any lengthly
>  discussion of his great efforts. :-) )


Thanks for the kind words John.

As well as a least squares curve fit to the historical data I've been using a form of variance analysis that looks at the block sizes being allocated by the registry over the past 12 months, and calculates the relative probability that on any day ARIN will allocate another block of that size. I run this model using a random seed to drive the probability factors and generate 100,000 individual outcomes, and plot the end data across all these runs of the model. The result is at http://www.potaroo.net/tools/ipv4/plotvar.png

Currently the data is pointing to ARIN exhausting its address pool in early March 2015, which coincides with another small round of IANA handout from their recovered pool. There is a small probability that the ARIN IPv4 pool will not last past December 2014, and equally a small probability that it will last until mid 2015, and the most likely outcome is runout around the March 2015 timeframe.

regards,

  Geoff