Hi Ran, On 26/03/2013 16:28, RJ Atkinson wrote: > Hi, > > The presentation on U/G bits that is archived online (PDF) > at the IETF web site has the text below on Slide 4: > > ILNP does have the constraint that its Node Identifiers > must be unique within a given site, but as we have just shown, > the state of the u bit does not in any way guarantee this. > > That statement quoted just above is NOT correct. > > The correct statement is: > > ILNP has the constraint that a given Node Identifier > MUST be unique within the context of a given Locator > (i.e. within a single given IPv6 subnetwork).
OK - thanks for that. > This is > guaranteed by existing IPv6 mechanisms (e.g. IPv6 ND). As far as I can see, ND+DAD guarantees that non-uniqueness will be detected, but does not guarantee that it will not occur. In any case, we certainly agree that a subnet with ambiguous identifiers won't work, which is the point. Brian -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------