> I stand corrected. That said, updating the specs to allow a site to use stateless address autoconfiguration with prefix lengths other than /64 would almost certainly require updating both specs.
The stateless autoconfig spec would need to be tweaked to convert the IID produced by the specific link-layer into something that is either extended (or shrunk) to match the advertised prefix in an RA. And we'd then have to recover from address collision (duplicate addresses), which the current spec does not, and generate (and try) new addresses, sort of like the temporary addresses spec defines. This could probably all be defined in a way that is an optional extension to stateless address autoconfig. So it wouldn't necessarily cause "confusion" or delay getting IPv6 deployed. However, I wonder what incentive device implementors would have to actually implement it. They seem (these days) to only implement the bare minimal functionality needed to get by. Extensions that might be useful in only some deployments, don't seem to be high priority. And, if you have a network in which 80% of the devices implement the new spec, but 20% do not, how useable would such a feature be in practice? And, keep in mind, one can use abitrary prefixes by turning off stateless address autoconfiguration and using DHCPv6. Thomas -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
