> > Some RFCs (I know of at least 2, RFC2526 and RFC4214) reserve a set > of > > interface identifiers on all prefixes. These identifiers need to be > > excluded when a node autoconfigures an address. This problem occurs > with > > privacy addresses but is equally applicable to other address > assigment > > methods like dhcpv6, cga etc.
Aren't we putting the cart before the horse? An IANA registry cannot be the only way to avoid collisions. An address allocation program may take into account the state of the registry at the time the code is written, but is seldom updated after that. Just getting a number allocated by IANA will not prevent competing use of that number for other purposes. There is a well established way to reserve MAC addresses, using a range of identifiers assigned to IANA. There is also a well established way to derive IPv6 interface identifiers from MAC addresses. It has been specified for over 10 years, and you can be sure that all IPv6 implementations abide by it. If RFC 2526 and RFC 4214 followed that framework, there would not be any problem. So, should we not instead go update these two RFC? -- Christian Huitema -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------