Re: [OPS-AREA] [OPSAWG] Heads up - Plea for allocating a /8 to ISPs
james woodyatt <[email protected]>
| Newsgroups | gmane.ietf.v6ops,gmane.ietf.ops |
|---|---|
| Message-ID | <[email protected]> |
On Nov 11, 2010, at 01:07, Christopher LILJENSTOLPE wrote: > On 11Nov2010, at 19.10, james woodyatt wrote: >> >> On 11/11/10 2:11 PM, "james woodyatt" <[email protected]> wrote: >>> >>> What advantage is there in directing *IANA* to do it? >> >> I understand why you want to use a prefix not listed in RFC 1918. I don't understand why getting it from IANA will be any better for you than if you get it from Joe's House Of IPv4 Addresses. > > It's simple, operators out of at least 2 of the 5 RIR regions are saying they need this. I would bet that operators in all RIR regions would use this if available (heads pulled in after the bruising battle this has been in the past). As it is "global" it does not seem fair to ask one RIR to sacrifice for the rest. If you are not talking RIR, then transfer policies make that pretty much a one-way street for the donor. The draft is asking for IETF to direct IANA to deny the NRO access to some currently unallocated prefix and declare it a new kind of special-use prefix. I wonder what the NRO thinks about that, given that their mission explicitly includes the protection of the free pool from barbarians. It's important, because you haven't made an *engineering* case yet for why this prefix needs to be a special-use reservation by IANA as opposed to an ordinary one allocated by the usual method to some organizational vehicle created for the purpose, i.e. The NAT444 Operators Group, Inc. Moving on. The draft uses RFC 2119 normative keywords despite being categorized as Informational. It places exactly one "requirement" on CPE implementations: "Network equipment manufacturers MUST NOT use the assigned block in default or example device configurations." This is a pointless requirement. There is no real danger of this ever happening with ordinary global addresses currently in the free pool. The rest of the "requirements" in the draft are equally pointless for the same reason. Moreover, Section 3 "Motivation" makes it clear that the operational consideration here is not about future CPE gateway devices anyway. It's about legacy devices, *none* of which use any of the address prefixes currently in the IANA free pool in default or example configurations. Therefore, the requirements do not follow from the motivation. I fail to see why NAT444 operators, as a group, need IETF to deny a perfectly good global IPv4 prefix to the NRO, when an ordinary global allocation to an assigned special entity would suffice for the purposes in question. -- james woodyatt <[email protected]> member of technical staff, communications engineering _______________________________________________ v6ops mailing list [email protected] https://www.ietf.org/mailman/listinfo/v6ops