[members-discuss] Input from Membership on RIPE NCC Charging Scheme Model

I don't think people really understand how RIPE works and what little 
power financially speaking they have.

Lets take this from RIPE's own annual report:

The RIPE NCC’s most prominent tasks include:
• Registering and distributing Internet number resources
• Operating the RIPE Database
• Operating K-root, one of the world’s 13 root name
server clusters
• Facilitating RIPE community activities
• Developing the RIPE Atlas network
• Providing high-quality measurement information
services

So yes, RIPE maintains a big database, that is their job.

Take time, read over their financial reports. RIPE has an operating cost 
of about 21,978,000 EUR per year with an income of 26,878 kEUR, so a 
refund/rebate of 5,335 kEUR was sent to everyone on their renewals the 
following year.
To date, RIPE has 14,344 LIR's , so back in 2015 based on the Annual 
report they had about 12,830 Paying LIR's.

Let's do the math here.

Based on the accounting and *cost* to run RIPE , for 12,830 LIR's the 
average cost per LIR would be 1,713 EUR / year. With even more LIR's, 
the cost to maintain them will not go up that much, so the cost per LIR 
has dropped even more.

If someone wants to get a total number of addresses managed by RIPE then 
based on the above numbers we can easily calculate down to a per IP charge.

But lets have some fun. 4,294,967,296 addresses are in the world, of 
which IANA has 16 /8's reserved for future use, and 16 more /8s reserved 
for multicast use only. Plus one more /8 held back for local 
identification. So lets erase 553,648,062 addresses from that number to 
start.

So down to 3,741,319,234 addresses to go. Minus another 92 /8's that are 
still in legacy status according to IANA

So we're down to 2,197,815,546 resources to split between RIPE, ARIN, 
APNIC & LANIC.

Let's make some assumptions now. Lets just say, since RIPE is a pretty 
big region, it gets 40% of those addresses. So 879,126,218~ addresses to 
play with. So that's 0.024 EUR PER IP PER YEAR. Making the cost of 
someone with a single /22 to be... 24 EUR per year.

Now, if someone had a /8 , sure they are going to pay a pretty good 
chunk. 402,653 EUR per year. But let's face it. You are going to 
encounter two distinct facts here.

1) The low cost of a /22 will exhaust resources immediately, as the 
total sum collected from signup fee's was refunded and reimbursed to 
members the following year, so the signup fee argument is invalid.

2) The guys with the /8 will guaranteed invest 400,000 EUR in their 
legal team, before paying RIPE 400,000 EUR per year per /8 which will 
inturn increase the cost of RIPEs operation (or just put it out of 
business completely). Then where will we be? Higher fee's to fight off 
the legal battles or wasted years and lots of refunds because of 
ill-advised changes.


Not to be an ass, but hopefully this will put an end to this silly 
discussion. The most fair way is a relatively equal playing field based 
on the LIR and not the amount of resources one maintains.

Keep in mind, the more IP's RIPE manages, the lower this cost goes. My 
879~ Million estimate is just that, a rough estimate I pulled out of 
relatively thin air. If someone wants to provide actual stats for how 
many IP's are in the RIPE pool, minus Legacy space which would be hard 
to bill please feel free to add to this.

Otherwise have a great day.





On 09/15/2016 06:41 AM, Brandon Butterworth wrote:
> On Thu Sep 15, 2016 at 08:48:59AM +0200, Bastien Schils wrote:
>> On 14/09/2016 16:24, Rob Golding wrote:
>>> On 2016-09-13 20:57, Chris Smith wrote:
>>>> It seems clear to me that LIR's that have large IPv4 resources are at
>>>> an economic advantage against LIR's with small IPv4 resources.
>>> Whilst that might be true in your business model, not everyone is
>>> using your model for their organisation
>> Can we keep it a clean conversation without "ad hominem" attacks?
>>
>> That remark was quite useless at best, disrespectful at worst.
> No it is not an attack, useless or disrespectful.
>
> It is central to the lack of consensus on charging model. LIR
> are a diverse community, ignoring that will only prolong the
> discussion
>
> The flat pricing is the simplest model to apply to all. If a
> proportional model is desired then it should relate to the cost
> of service not the number of IPs or prefixes unless they have
> a direct cost (they were allocated to RIPE for free?)
>
> As a small LIR we would benefit if it was based on total IP
> count but be worse off if it was quantity of prefixes as
> we've had a number of small allocations due to not obtaining
> a single large one early.
>
> We'd probably benefit if it was based on hours of support used
> as we interact with RIPE very little, usually initiated by RIPE
> as some voted to change something.
>
> I'm not sure why RIPE initiated this discussion, it just rakes
> over old arguments that had no route to consensus
>
> brandon
>
> ----
> If you don't want to receive emails from the RIPE NCC members-discuss
> mailing list, please log in to your LIR Portal account and go to the 
> general page:
> https://lirportal.ripe.net/general/
>
> Click on "Edit my LIR details", under "Subscribed Mailing Lists". From 
> here, you can add or remove addresses.




On 09/15/2016 06:13 AM, Floris Bos wrote:
> On 09/15/2016 09:46 AM, Carlos Friacas wrote:
>>>> >  If another LIR has a hundred times more IPv4 addresses than we do,
>>>> >  then I'd expect them to pay 100 times (or more) than we do.
>>>>
>>>>  And therein lies the difference in thinking - if one LIR uses 100
>>>>  times the "resources" than another then yes, a larger bill could be
>>>>  appropriate. But a range of ips is ultimatley just "1 resource" - it
>>>>  doesn't matter about the size of that range.
>>> Let's take it from another perspective:
>>> If a LIR has 100 more IPs than another, wouldn't it be expected to 
>>> think that this LIR is 100 times more likely to need actions from 
>>> the RIPE?
>>>
>>> Well, in my opinion: Yes.
>>
>> I'm not sure about that..... newcomers (which by the current policy 
>> only get a /22) are probably where the NCC is spending the largest 
>> part of their effort. Is there any measurement already?
>
> Would argue that any extra effort is already paid by the newcomers 
> themselves in the form of the 2000 EUR setup fee.
>
> And as a recent newcomer I am not really that impressed by what you 
> get for that amount.
> If you are a Dutch company they do not even send the paperwork by 
> trackable means, but do not want to spend more than a stamp on you.
> The first time they send the agreement by regular postal mail it never 
> arrived, and it took three weeks before they could be bothered to 
> resend it a second time.
> That did arrive, despite them putting the address in the wrong format 
> for my country on it, which doesn't exactly speed up postal sorting 
> and delivery either...
>
>
> Anyway, back on topic.
> I would argue that differentiating membership fees based on size of 
> assignment do would be fair.
> The cost per customer is a lot higher for smaller members than large 
> ones, and as such the latter have an economic advantage.
> Furthermore the setup fee already provides a sufficient barrier 
> preventing the number of companies signing up as LIR to explode.
> No need to keep the membership fee artificially high for smaller 
> member for that purpose.
>
>
> Yours sincerely,
>
> Floris Bos
>
>
> ----
> If you don't want to receive emails from the RIPE NCC members-discuss
> mailing list, please log in to your LIR Portal account and go to the 
> general page:
> https://lirportal.ripe.net/general/
>
> Click on "Edit my LIR details", under "Subscribed Mailing Lists". From 
> here, you can add or remove addresses.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.ripe.net/ripe/mail/archives/members-discuss/attachments/20160915/196a4401/attachment.html>