Hi, I also oppose the policy and see no sense in coming with such proposal today. I have seen many good points and I would like to express my opinions on the pros/cons arguments from the proposal: PROs: - The RIPE NCC has the strictest policy of allocations from the remaining IPv4 address pool. All of the other RIRs allow more than a single /22 per member as long as enough address space is available. I would not care of what happens in other regions and the "enough" is always relative. -36 months after the activation of the “last /8 policy”, the RIPE NCC still had more than 99% of the equivalent of a /8 mostly due to address space being recovered and re-allocated by IANA. Exactly, if IANA would not have recover and re-allocate IPv4 resources to RIPE, then almost half of the last /8 would have been gone. There's no guarantee that IANA will keep supply RIPE's available pool so the 99% figure is totally wrong to be used in this context. - In a number of markets, not being able to provide at least a single IPv4 address per customer prevents members from performing, regardless of the ample availability of IPv6 address space. The large telcos need to perform and they need huge IPv4 resources, should we be concerned of their performance ? - Faster depletion of the free IPv4 pool may force the adoption of IPv6 on certain members. Not really. As we have seen, every country that has relatively large IPv6 adoption rate it's just because of a single dominant provider, the move to IPv6 will be driven by large telcos not the small ones and they would not be influenced in any way by this policy change. - LIRs that opened after 14 September 2012 can only have a /22 allocated by the RIPE NCC as per the current policy, which gives them a disadvantage compared to older LIRs that were able to obtain at least a /21 regardless of their needs. Here I agree, the last /8 policy should have given a /21 to each LIR, but that should have been done at that time based on the fact that previously a /21 was allocated by default and it would have been fair to use the same "unit". However if IANA would not have given RIPE IPs, today we would have exhausted the last /8 (by allocating /21s). So I think giving a /22 was a very good idea at the time and doubling today the allocation would make us hit the bottom probably in less than a couple of years. We have proven that we are unable to predict the future. So, let's not suppose that things will be nice and IPv4 will not be needed in 5 years and let's leave theese drops of milk for the future children. As I mentioned I don't agree that hitting the bottom of last /8 pool will help IPv6 deployment. It's driven by the large telcos and if a new entrant would not be able to get a few IPv4 resources from RIPE, they would only have to buy from the ones that sit on the resources so it can only help prices to increase and make it harder for a new entrant to start a business. Ciprian On 10/20/2015 9:31 PM, Jim Reid wrote:
This policy proposal is a Very Bad Idea. I oppose it.
The arguments presented in support of this proposal make no sense. IMO 2015-05 is harmful and not in the best short- or long-term interests of the RIPE community.
There may be a case for making an LIR's final IPv4 allocation larger than a /22. However that position is not yet backed up by any supporting data and would need to go in another policy proposal anyway.
I hope the WG kills 2015-05.