Re: [address-policy-wg] 2012-04 New Draft Document Published (PI Assignments from the last /8)
On Mon, Sep 10, 2012 at 02:16:44PM +0200, Emilio Madaio wrote:
I do fully support this proposal. Thanks Nick for bringing up that issue which I missed when the last-/8 policy proposal was on the table. To those who ask for "same price tag" for both PA and PI on grounds of "fairness"... What is fair in asking the same price for what is essentially a one-time operation (PI assignment) compared to the ongoing maintenance of a LIR membership (PA alloc - think about assignment request tickets, yearly billing, etc.)? Best regards, Daniel -- CLUE-RIPE -- Jabber: dr@cluenet.de -- dr@IRCnet -- PGP: 0xA85C8AA0
* Daniel Roesen
To those who ask for "same price tag" for both PA and PI on grounds of "fairness"... What is fair in asking the same price for what is essentially a one-time operation (PI assignment) compared to the ongoing maintenance of a LIR membership (PA alloc - think about assignment request tickets, yearly billing, etc.)?
The comparsion I'm making is between an organistion that either: * Joins the NCC - becomes a LIR * Makes 1 initial allocation request - gets a /22 * Makes 1 PA assignment request (to itself) Cost: €1750 Or: * Makes 1 PI assignment request [through a sponsoring LIR] Cost: €50 I don't believe that the current prices are «fair» either. While the first option is certainly means work for the NCC, I doubt it is 35 times the work. The resource is scarce - it is impossible to accomplish «fair». In any case, the reason why I'd like PI and PA to cost about the same is that 2012-04 with the current pricing structure would mean it is too easy and cheap for an organisation to work around the «max a /24» limitation. I could trivially set up legal entities like so: * ToreISP Main Street 1-80 Oslo * ToreISP Main Street 81-120 Oslo * ToreISP Grand Avenue 20-35 Stockholm ...and so on, and get all the PI /24s I need. The organisations that are actually NCC members will end up sponsoring me, too - potentially ending up with even higher membership fees, as the Impact Analysis points out. It may be circumenting policy, but I don't think there's any way to stop it. (The NCC mentions this in the Impact Analysis, too.) That's why I think there must be a financial disincentive against doing so, and €50 doesn't cut it. I'll end with an example of something similar already taking place: Under .no, a single organisation can only register 100 domains. Which leads to organisations such as these: http://w2.brreg.no/enhet/sok/treffliste.jsp?navn=edda+domene This isn't the work of a shady domain pirate/squatter either, Edda a very large and respected media/publishing conglomerate here. -- Tore Anderson Redpill Linpro AS - http://www.redpill-linpro.com
On Tue, Sep 11, 2012 at 10:52:10AM +0200, Tore Anderson wrote:
* Joins the NCC - becomes a LIR * Makes 1 initial allocation request - gets a /22 * Makes 1 PA assignment request (to itself)
Cost: ?1750
Or:
* Makes 1 PI assignment request [through a sponsoring LIR]
Cost: ?50
But they can do this now. In fact it's always been possible to game that system, regardless of ipv4 run-out. Assuming this is not a widespread problem now, why would it become one? rgds, Sascha Luck
* Sascha Luck
On Tue, Sep 11, 2012 at 10:52:10AM +0200, Tore Anderson wrote:
* Joins the NCC - becomes a LIR * Makes 1 initial allocation request - gets a /22 * Makes 1 PA assignment request (to itself)
Cost: ?1750
Or:
* Makes 1 PI assignment request [through a sponsoring LIR]
Cost: ?50
But they can do this now. In fact it's always been possible to game that system, regardless of ipv4 run-out. Assuming this is not a widespread problem now, why would it become one?
Precisely because of IPv4 depletion. Say you're an ISP who needs a /16. Today, you can have it - and there's no reason to game the system. However, after the last /8 policy hits, you can't get this space from the NCC unless you either (assuming 2012-04 passes): A) join the NCC with 64 LIRs, each getting a /22 allocation at a total cost of €112.000, or B) get 256 /24 PI assignments at a price of €12.800. The B option is actually affordable. At €0.20 per address, it's way cheaper than what Microsoft paid for the second-hand Nortel addresses, for instance (US$11.25 per address IIRC). -- Tore Anderson Redpill Linpro AS - http://www.redpill-linpro.com
On 11 September 2012 12:02, Tore Anderson <tore.anderson@redpill-linpro.com> wrote:
A) join the NCC with 64 LIRs, each getting a /22 allocation at a total cost of €112.000, or B) get 256 /24 PI assignments at a price of €12.800.
The B option is actually affordable. At €0.20 per address, it's way cheaper than what Microsoft paid for the second-hand Nortel addresses, for instance (US$11.25 per address IIRC).
and option A is less than e2.00 per year so excellent value for money if you compare it to the MS price J -- James Blessing +44 7989 039 476 Strategic Relations Manager, EMEA Limelight Networks
* James Blessing
On 11 September 2012 12:02, Tore Anderson <tore.anderson@redpill-linpro.com> wrote:
A) join the NCC with 64 LIRs, each getting a /22 allocation at a total cost of €112.000, or B) get 256 /24 PI assignments at a price of €12.800.
The B option is actually affordable. At €0.20 per address, it's way cheaper than what Microsoft paid for the second-hand Nortel addresses, for instance (US$11.25 per address IIRC).
and option A is less than e2.00 per year so excellent value for money if you compare it to the MS price
Absolutely. I don't dispute that this can happen the «multiple LIRs» way, regardless of 2012-04. My worry is mainly that 2012-04, when combined with the current pricing model, will greatly exacerbate the problem. -- Tore Anderson Redpill Linpro AS - http://www.redpill-linpro.com
You would need to have 256 companies for this. This, in turn, generates significant cost when compared to 64 companies. 64 LIRs give voting power, as well. Is this scenario realistic? Sent by mobile; excuse my brevity.
On Tue, Sep 11, 2012 at 1:26 PM, Richard Hartmann < richih.mailinglist@gmail.com> wrote:
You would need to have 256 companies for this. This, in turn, generates significant cost when compared to 64 companies. 64 LIRs give voting power, as well.
Is this scenario realistic?
Yes, as Tore already explained, this strategy has already been successfully used for registering .no domains beyond the registry imposed limit. Registering a .no domain costs merely 20€ or so, and until this year, there was a cap at 20 registrations per organization. Feel free to ask the owners of Domene Klubben 1 through 442 if the inconvenience of registering 441 extra organizations was too much, even with all the manual paperwork needed back when they formed in 2000. In the case of IP addresses, the cost savings are significant enough that the bother of registering a few hundred extra organizations shouldn't be too bad. -- Jan
* Richard Hartmann
You would need to have 256 companies for this. This, in turn, generates significant cost when compared to 64 companies. 64 LIRs give voting power, as well.
Is this scenario realistic?
I suppose it depends on how complicated it is to register a legal entity, which probably varies greatly from country to country. In Norway, it is very simple and to the best of my knowledge does not cost anything. Or, as the NCC puts it: «It is also relatively straightforward in many jurisdictions to set up a separate legal entity that could then apply for a PI assignment, resulting in one actual organisation holding multiple /24s from the last /8.» Considering that registering multiple legal entities to circumvent restrictions on how many .no domain names a single one can hold is already happening -- yes, I do think it is a realistic scenario for acquiring IPv4 as well (perhaps even more so). Registering 192 extra legal entities in order to save €100K is a no-brainer, if you ask me - even when taking into account the lost voting rights. I'm guessing that an organisation that would do something like is motivated by desperation for IPv4 addresses, rather than a desire to participate in the RIPE community or NCC membership. -- Tore Anderson Redpill Linpro AS - http://www.redpill-linpro.com
participants (6)
-
Daniel Roesen -
James Blessing -
Jan Ingvoldstad -
Richard Hartmann -
Sascha Luck -
Tore Anderson