[address-policy-wg] 2019-02 New Policy Proposal (Reducing IPv4 Allocations to a /24)
- Previous message (by thread): [address-policy-wg] 2019-02 New Policy Proposal (Reducing IPv4 Allocations to a /24)
- Next message (by thread): [address-policy-wg] 2019-02 New Policy Proposal (Reducing IPv4 Allocations to a /24)
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Carlos Friaças
cfriacas at fccn.pt
Tue Feb 5 08:58:42 CET 2019
On Tue, 5 Feb 2019, Tore Anderson wrote: > * Marco Schmidt > >> A new RIPE Policy proposal, 2019-02, "Reducing IPv4 Allocations to a /24" >> is now available for discussion. >> >> This proposal aims to reduce the IPv4 allocation size to a /24 once the >> RIPE NCC is unable to allocate contiguous /22 ranges. > > I support this proposal. Some random thoughts about it: Hi, Some comments inline: > - It is good to have the policy explicitly say there is a waiting list. > Current policy says nothing about this. If the NCC would simply close > allocation tickets with «sorry, fresh out, try later», that could > encourage LIRs to spam repeated allocation requests in the hope that > theirs was the first request to be received after an IPv4 fragment > was returned to the allocation pool. Yes, and that "randomness" would be everything except "fair". > - I don't quite believe that the waiting list would grow indefinitely > (regardless of the allocation size being /22 or /24). Keep in mind > that only new and IPv4-less LIRs would be able to join the waiting > list. Once it is known that simply joining the NCC won't guarantee > a /22, but that you'd have to wait for one with no certainty as to > how long, I expect that the sign-up rate of new LIRs will drop > dramatically (and by extension the amount of LIRs queuing up in the > waiting list). If getting a /24 is still cheaper than getting a /24 from "the market", people will queue up, because there will be still a bit of profit... On a personal viewpoint, i also hope newcomers come to the RIPE NCC for IPv4, so they can be flooded with information about IPv6 :-) > - It seems reasonable to lower the allocation unit to the de facto > smallest usable on the public Internet at a point in time when we > can no longer allocate /22s (which are already pretty small). > Otherwise recovered /23 and /24 fragments (e.g., PI assignments) > will just end up rotting away in the NCC inventory, which serves > no good purpose at all. Yes, growing up the NCC's IPv4 inventory will serve noone, except if a policy is accepted in the future to re-open IPv4 distribution, when the inventory reaches some level. I clearly prefer to see /24s distributed to those who want them. > - It seems reasonable to trigger this policy at precisely at a > watershed moment like the policy aims to do (unlike 2017-03, which > would have changed the rules in the middle of the game). As you may have noticed i was also one of the co-authors of 2017-03, and that one was withdrawn. But the current proposal is not 2017-03, and i feel this is needed especially after getting input from the NCC about the amount of address space they are getting back -- due to several reasons. > - The authors should clarify how this new policy interacts with the > /16 set aside in «5.2 Unforeseen circumstances». In which order > does the 5.2 and 5.1bis policies get triggered? I wouldn't touch that. Would let the NCC decide what are "unforeseen circumstances" or wait for a new, different, policy proposal. We might tackle this issue at v2.0, but i would like to keep changes at a minimum, in this proposal's scope. Best Regards, Carlos > Tore >
- Previous message (by thread): [address-policy-wg] 2019-02 New Policy Proposal (Reducing IPv4 Allocations to a /24)
- Next message (by thread): [address-policy-wg] 2019-02 New Policy Proposal (Reducing IPv4 Allocations to a /24)
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]