[arin-ppml] Call for a study & survey to obtain necessary information for policy development

[Michel Py]

Jeffrey,

> Jeffrey Lyon wrote:
> where other companies who managed to "justify" /17
> and shorter are freely handing out space

This is a very common method of stockpiling, unfortunately. Recently, I
set up a small network for a small customer (5 sites), all in PA space,
a mix of fiber, T1 and even sDSL. The customer negociated with the ISP,
all I asked was 1 static IP per site and 5 for the main one. Well, I got
handed a mix of /29, /28 and /27 and while I actually configured a total
of 7 IP addresses, there are well over 100 of them wasted. It does not
take a rocket scientist to figure out that when the ISP involved starts
to run out of addresses, they are going to start charging for them and
renumber customers so for them the shortage is a joke. We know that some
companies are hoarding resources either in hopes of future resale or to
cover an arbitrary future need.


Which brings me to this:

> - In the survey, ask IPv4 resource holders to anonymously disclose 
> their true utilization rates and determine if companies are hoarding 
> resources either in hopes of future resale or to cover an arbitrary 
> future need.

I don't think that this one item is worth spending time and money on.

First, we would have to deal with {I don't have time for this | I don't
trust you | I don't know | I don't care | etc}. I am uncertain that data
collected would be usable at all.

Second, even if what I wrote just above did not exist, the utilization
numbers that you will get are not real, as they include the hoarding
factor. Related to the first part of my post, the data collected would
show 100+ addresses in use, while the reality is 7. What method do you
propose to separate the hoarding from the real in the data set?

Finally, what would you do with that data if you had it?


> Determine if resource holders would be encouraged to tighten
> up internal policies and free up more space if there were a
> fair market value assigned to their space.

I think this question is meaningless until you put a figure on it.
At $1000 per IP? Heck yes. At $1 per IP? Heck no.

For the record, I agree with most of what you wrote recently.

Michel.