Donald Eastlake <[email protected]> writes: > IPv6 Interface Identifiers (IIDs) are 64 bit quantities. Various RFCs > say you can use a Modified EUI-64 for an IID on the basis that you > control the base EUI-64. And control of some EUI-64s can derive from > control of an EUI-48 (i.e., a MAC address) or IPv4 address. It would > seem unfortunate if someone specified some IID for special use when > someone else could start innocently using it, believing it was theirs, > due to their control of an EUI-64 or EUI-48 or IPv4 address.
IPv6 address with modified EUI-64 IIDs have (in some cases) special properties. Maybe. But not all of them do, depending on the u/g bit setting. To date, no one really takes advantage of those properties. The original purpose for them was always speculative, and in the 15 years since then, I've only become less convinced that those properties are useful in practice. For one thing, because there is no way to force the properties to be valid, any implementation/use that depended on such properties would end up being a potential denial of service vector from Bad Guys that deliberately generate IIDs in violation of the assumption. Also, it was always assumed that these properties might not actually be available in practice, depending on what implementations actually did, etc. Given that, I don't see the point in worrying a whole lot about EUI-64 IIDs or expending much effort on them. Not unless we have a credible proposed usage for them. Thomas -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
