[address-policy-wg] Update on ALLOCATED PI/UNSPECIFIED
- Previous message (by thread): [address-policy-wg] Update on ALLOCATED PI/UNSPECIFIED
- Next message (by thread): [address-policy-wg] Update on ALLOCATED PI/UNSPECIFIED
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Gert Doering
gert at space.net
Thu Aug 4 14:10:33 CEST 2016
Hi, On Thu, Aug 04, 2016 at 08:59:07PM +0900, Randy Bush wrote: > > Right now, there are two different shades of "PI colour" - "real PI" > > and "not really real PI". > is there a list of all the colors and what they mean? I used to assume there is "ALLOCATED PA", "ASSIGNED PA" and "ASSIGNED PI", and those are well-defined. Add "Legacy" to it (outside RIR framework). We have learned that there are more interesting shades, namely "ALLOCATED PI", "ALLOCATED UNSPECIFIED" and "ASSIGNED PI (not really)" - these are not seen often, and their meaning is not all that clear. [..] > i have this feeling you are trying to say something here. i.e. if i am > the LIR, can i move "not really real PI" between customers and no one > knows? Maybe. "not really real PI" is happening outside the RIPE policy framework - namely, no 2007-01 contracts, the NCC has no idea who the "real" address holder is, and the holder cannot complain to the NCC if all of a sudden "his" address space is registered to someone else... There seems to be "old stuff that happened 15+ years ago" and "new stuff that is still ongoing", so there won't be "one solution" either. > > Also, it might lead to deaggs (Markus' case) where a /14 that was > > originally "in one LIR" would be "3x /16, plus some smaller fragments > > in the LIR" and "lots of /24 PI managed by the NCC" now - so the /14 > > won't get a ROA, and he'll have to announce more-specifics. > > lemme see if i get this. to have the owner registration correct, some > address space will have to be broken up and owned by multiple IRs, thus > fragmenting routing? i like correct registration, but the commons has > become pretty polluted. I leave the definite answer to Ingrid to answer. My understanding of "normal" NCC<->LIR stockkeeping is that PI is never living inside blocks that "belong" to a given LIR. So, the LIR would never be able to get a ROA covering PI space. For some of these "old" blocks, there is a /16 which covers regular LIR/PA space, and "not real PI" space, and the LIR can get a ROA that covers their PA space, and also these "not real PI" blocks (because according to the NCC records, the /16 "belongs" to the LIR). From an aggregation PoV, this is ok-ish - but from a routing security PoV, I wonder if that's what we want (the "not real PI" block might be routed totally elsewhere now). > > So, to answer your question: for those "swampy PI", it would alter > > their rights (contracts according to 2007-01), costs (50 EUR/year) > > whoops. that's gonna cause unhappiness. Dunno. We (the RIPE community and the NCC) rolled out 2007-01 to all the other PI holders, and the amount of unhappiness was not very big. Those cases that I was involved with my "LIR admin-c" hat on, PI holders seemed to be happy to have a clear contract with a known entity (us), and the assurance that this would ensure that nobody else could make claims to their address space. Gert Doering -- assorted hats -- have you enabled IPv6 on something today...? SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (0)89/32356-444 USt-IdNr.: DE813185279 -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: not available URL: <https://lists.ripe.net/ripe/mail/archives/address-policy-wg/attachments/20160804/b239865c/attachment.sig>
- Previous message (by thread): [address-policy-wg] Update on ALLOCATED PI/UNSPECIFIED
- Next message (by thread): [address-policy-wg] Update on ALLOCATED PI/UNSPECIFIED
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]