[arin-discuss] [arin-ppml] Fee proposal (was Re: Alternativetoarbitrarytransfers)

[Joe Maimon]

Alexander, Daniel wrote:
> Ted,
resources is not because they are cheap, or easy to come by. It is 
because they are connecting end users (I assume you are one of them) to 
the Internet. The growth, and very existence of an ISP is based on the 
services they provide and the revenue these services generate.
> 
> The fact that IPv4 is running out is the single largest incentive any ISP needs to deploy IPv6. 

So why hasnt it happened yet?

> and are more often passed along to the consumer. 

Consider that currently the fee for ip address passed on to the consumer 
is effectively zero, that would actually be good thing.

Compare the annual cost to a provider of an end-user /29 (Leased line, 
non residential) and per IP (residental/soho) based upon ARIN's current 
fee schedule. Rounded.

X-Small 1250 per /21 = 4.8 (0.6)
Small 2250 per /19 or /20 = 2.2 or 4.4 (.28 or .54)
Medium 4500 per /16 /17 /18 = .54 1.09 2.19 (.07 .14 .28)
Large 9000 per /14 /15 = .27 .55 (.03 .07)
X-Large 18000 /8 /9 /10 /11 /12 /13 = .01 .02 .03 .07 .14 .27 (.001 .002 
.004 .01 .02 .03)

When sprint or cogent or att signs up a T1 customer, the customers /29 
costs them as little as one cent annually. If the customer demands a 
/24, they barely care.

When any ISP grossing under a few million a year signs up a T1 customer, 
  the /29 can cost them up to 5 bucks. A /24 is a serious customer 
relation issue.

This is why for the history of the internet, end users got their space 
free, because a small provider cannot compete by charging fees large 
providers dont even bother thinking about.

The big boys have been getting quite the free ride for the entire 
history of the net - and they are the ones using up all the addresses.

Who do you think is calling the shots here?

Quite possibly they are the ones inefficiently wasting them as well.

Which network has the incentive to track their usage tightly? To be 
strict with allocations? To be inventive in their utilization so as to 
minimize the unnecessary?

Not the one where each address cost less than a penny.

And the more you waste now, the more you will be able to scavenge later.

Why do you expect the end user to want ipv6? Ipv4 is great and costs 
them nothing. This will need to change.

Of course, I am omitting the overhead costs, but those are a lot harder 
to quantify and their impact may be negligible.

When runout happens, assuming a provider cannot get any more ipv4 for 
whatever reason, they will scavenge internally. Then they will charge 
per ip address, incentivizing their user-base to go ipv6-only.

Joe