Le 2012-12-13 à 09:13, Brian E Carpenter <brian.e.carpen...@gmail.com> a écrit :
> On 12/12/2012 21:32, Ole Trøan wrote: >>> I think the only way to make progress on this question is to >>> discuss two points in a much more general way: >>> >>> 1. Does the current mapping of the "u" bit in the modified EUI format have >>> any value? >>> >>> 2. Does the current mapping of the "g" bit in the modified EUI format have >>> any value? >>> >>> If the answer to these questions is "no" we could start by saying that they >>> are *not* mapped from an IEEE EUI into a modified EUI, and then consider >>> how to use them in future. >> >> or we could consider the interface-id as 64 bits without any implied meaning >> attached to >> any of the bits. > > We could certainly discuss that, but first I'd like to understand if the > mapping > of the u/g bits into layer 3 addresses has any value in practice. > > I believe it's true that the u bit has no value for ILNP. At one point there > was > a thought that it had potential value for 8+8. Maybe Ran Atkinson can comment. > > Does the g bit have any value in multicast? (Obviously it has value in the > MAC address, > but here we are talking about the IP address.) IPv6 unicast addresses having the group bit set to 1 *already exist* (with the privacy option where all bits other than u=0 can take any values). It is therefore ONLY the IPv6-address multicast prefix that determines that an IPv6 address is multicast. (In my understanding, this doesn't depend on any original thoughts about possible other uses of the g bit.) Another point is that, in RFC 4291, interface IDs are introduced ONLY for unicast addresses. For multicast addresses, what is introduced is Group IDs. IPv6 group IDs have no link, in any RFC that I know, with IIDs. (And even less with modified EUI-64). RD > > Brian > -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
