On Apr 14, 2012, at 5:07 PM, Christian Huitema wrote: >> As I read it, section three item one calls for the use of the EUI-64 in use >> on the interface, which presumes that the >> interface is an IEEE 802 LAN. There are other interface types. I'd like to >> see that widened to a number *such*as* >> one of the set I specified. > > You don't really need the EUI64 for the proposal to work. We could just as > well completely omit the field, and simply use the EUI64 as an optional seed > for the random number. But there are two advantages: > > * Using some form of interface ID allows the host to keep just one random > seed for all interfaces, which is nice. > * using EUI64 in a context like SEND makes spoofing a little bit harder.
I have no problem with the EUI-64 if one exists. I'm pointing out that not all interfaces have one. They might have an E.164 or E.212 number, or other things. -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------