[arin-discuss] [arin-ppml] Alternative to arbitrary transfers

Ted Mittelstaedt tedm at ipinc.net
Mon Apr 6 18:40:04 EDT 2009


 

> -----Original Message-----
> From: arin-ppml-bounces at arin.net 
> [mailto:arin-ppml-bounces at arin.net] On Behalf Of Brian Johnson
> Sent: Monday, April 06, 2009 2:36 PM
> To: ARIN PPML
> Subject: Re: [arin-ppml] Alternative to arbitrary transfers
> 
> > -----Original Message-----
> > From: arin-ppml-bounces at arin.net [mailto:arin-ppml-bounces at arin.net]
> On
> > Behalf Of Ted Mittelstaedt
> > Sent: Monday, April 06, 2009 4:09 PM
> > To: 'Kevin Kargel'; 'Leo Vegoda'; 'ARIN PPML'
> > Subject: Re: [arin-ppml] Alternative to arbitrary transfers
> > 
> > 
> <snip>
> > 
> > Unfortunately, many of these are operational, not policy, issues.
> > Particularly the fee proposals, an issue near to my heart 
> (as I don't 
> > work at a large ISP who gets the "Costco discount" on IPv4)
> > 
> > Ted
> 
> I'm not sure what you mean by Costco discount. 

Costco is a retailer known for it's practice of selling
large quantities of items at a cheaper per-unit price.

You don't go to Costco to buy a loaf of bread for $2.  They will
sell you 6 loaves for $9 though.

http://en.wikipedia.org/wiki/Costco

> If you mean 
> that there should be a direct cost per IP address that should 
> remain consistent across total assignment sizes, then I 
> understand your point.
> 

Yes.

> I always find these types of comments to be a bit grating and 
> tend to diminish the dialog instead of promote discussion. 
> There is no place in the policy that gives anyone a discount. 
> If you have more assignments, you pay more than others who 
> have fewer assignments. This is not a discount.
> 

Yes it is.  If you get more IPv4 assigned you pay less per IP.
This is a volume discount in every possible interpretation of the
phrase.

> If viewed by cost/IP, then the cost/IP for larger orgs (ISPs generally
> speaking) is less than for smaller orgs. This has been long 
> standing policy. If you want to change this. Make a proposal 
> and get consensus.
> Don't degrade one group to make yourself feel better.
> 

Volume discounts are a legitimate business practice and are not degrading!

In business they are used to do several things, protect the dealer
networks, protect the retail networks, and as an acknowledgement
that it's cheaper for the manufacturer to sell the same amount to 1 customer
than 100 customers

ARIN IP number discounts exist for many of the same reasons,
it protects ISPs  (do you really want all of your customers with
their own portable /29's) and because it is cheaper for ARIN to track
a SINGLE org that has a LARGE
assignment than MANY orgs that have SMALL assignments.  It is ALSO
CHEAPER for the REST of the Internet to route FEWER advertisements
from LARGER orgs than MANY assignments from SMALLER orgs.

My point here which I'm spelling out is that post-IPv4 runout, we
now have an IPv4->IPv6 conversion cost that must be bourne by
all networks.

Thus a large reason for the multiple-IP discount - that it's cheaper
to have larger orgs advertising fewer prefixes - is now counterbalanced
by the increased cost to switch to IPv6.  In other words, if we can
push off the IPv4 runout date by getting IPv4-heavy orgs to shed
some IPv4, that it will allow existing IPv4 orgs to stretch their old
equipment for a longer lifespan, thus saving the entire Internet money.

Plus, if we push-out the IPv4 runout date, then when runout happens
there will be a much more rapid "snap" to IPv6 - as conversion then will
be an absolute necessity because then no IPv4 will be available for
love or money.

This reduces the time that the entire Internet must dual-stack to
keep the laggards happy, and saves even more money for everyone.
It also drops the amount of data ARIN has to track, decreasing
their costs and allowing fees to drop.

It's pretty clear to me that cost/IP adjustments are a part of
any end-game IPv4 runout policy.  Of course, I do realize that
IP costs would have to be raised to obscene levels before the large
orgs would be motivated to shed IPv4 - but I also recognize that
the cost structure right now is seriously out of wack and very
unfair to the smaller players.  ARIN is also mandated to collect
enough fees to maintain it's operation - and having ARIN go on
fishing trips to pull dusty old IPv4 out of corners is going to
increase their expenses which will have to be paid somehow.  The
same exact thing would happen for the pro-transfer market proponents
too - having ARIN referee a bunch of paid transfers is also going
to increase their expenses.

And, who do you think will be the main beneficiaries of IPv4 that
is returned?  It won't be the small orgs who send in requests
once in a blue moon - it will be the large orgs who are chewing
up the IPv4 at a huge rate.

I may have personal dogs in this hunt, but I didn't bring this
issue up for myself.  After all, I don't own my employer - I
don't pay the yearly registration renewal fees myself.

Ted




More information about the ARIN-discuss mailing list