Hi Tore, On 6/9/2015 10:59 PM, Tore Anderson wrote:
Hi Ciprian,
* Ciprian Nica
What should be pointed out is the effects of the policy and if the community will benefit from it or some small group of people.
To summarize the effects will be : - higher membership fees
Nope. The RIPE NCC membership is steadily growing[1], and as a result the membership fee has steadily been decreasing[2].
[1] https://labs.ripe.net/statistics/number-of-lirs [2] https://www.ripe.net/publications/docs/ripe-620
The LIRs from Russia are 2nd after UK : http://www.ip-broker.uk/index.php?page=statistics.php&chart=lirs
The main reason for this growth is *actual network operators* joining in order to make use of the «last /8 policy». Even if we managed to stop *all* the "create LIR; transfer /22; close LIR" abuse, that would not reverse this trend. Also, keep in mind that these "create; transfer; close" LIRs will pay the NCC as little as they can get away with. As I understand it, that means the sign-up fee and one yearly membership fee. If the goal is to increase the NCC's revenue and lower the membership fees, it is much better long-term strategy to deny these "create; transfer; close" LIRs and instead keep the /22s in reserve for future LIRs belonging to *actual network operators*. Why? Because these will actually *keep paying their membership fees* instead of closing down as soon as possible.
I didn't say that I can measure how much it will affect but on the short term it will affect for sure the membership fees. Every newcomer will become an LIR and pay the fees so that won't change anything whether the abusers exist or not. For example the 2 mentioned entities have registered 61 new LIRs and sold the IPs. Let's say that paid 2000 signup fee and this year's fee in full which would mean a total 219.600 EUR. That probably covers the costs for organizing at least a few RIPE meetings. If the abusers would not have sold the IPs, the buyers would have got them from someone that received the IPs for free, maybe in 2012 and maybe more expensive. So I stick to my opinion that the membership fees will be affected.
What is the expected positive effect ? To preserve the last /8 pool ? The one that increased to 18.1 million IPs ?
The by far biggest contributor to the RIPE NCC's «last /8» pool has been the IANA IPv4 Recovered Address Space pool[4].
[4] https://www.iana.org/assignments/ipv4-recovered-address-space/ipv4-recovered...
This pool contained quite a bit of space when it was first activate, and the RIPE NCC has to date received 3,670,016 IPv4 addresses from it (/11+/12+/13). It is important to note, though, that the IANA pool *is not replenishing*. It has been almost three years ago since any significant amounts of space was added to it (back in 2012-08).
So we cannot expect that allocations from the IANA pool will continue to match the rate of /22 allocations from the RIPE NCC's «last /8» pool in the future. Therefore I have every expectation that we'll start seeing «last /8» pool actually start to drain soon.
For what it's worth, since the first «last /8» allocation was made 995 days ago (cake in five days!), a total of 6,657,280 IPv4 addresses has been delegated by the NCC. Our share of the remaining IANA pool is on the other hand only 425,625 addresses.
So all in all, I think that preserving the last /8 pool is indeed a valuable goal. If possible I'd like to see it last for another ten years - but given today's burn rate, the current 18.1M addresses plus whatever's coming from IANA will not suffice.
There is another source for IPs, the ones that were returned to RIPE either by LIRs that went bankrupt or by end users that didn't pay the maintenance fee or also went bankrupt. From what I remember there were quite a few PIs recovered by RIPE after the independent resources tax was introduced. Ciprian