Follow up: >To my understanding, rfc4941 meant to use CGA exactky as defined in rfc 3972.
To clarify what my purpose is, I want to address the situation where the node is not using stable storage but still needs to generate a random IID. This document is only an update to that section to provide a better approach when there is no need to rely on stable storage for randomization. >The modified CGA algorithm in draft-rafiee-6man-ra-privacy has nothing to do with CGA. The steps are the same except for removing the steps associated with security verification. I did not see any need for that. Finally, the modifier depicted in RA-privacy is just a random number. You can call it whatever you like, but it is not something secret that would compromise privacy if somebody knew it. The node picks up whatever number it likes. By combining this number with the timestamp and the subnet prefix a unique IID is generated. If another node would happen to choose the same number (that the probability of that is very low), then the timestamp enables the generation of a different value than that for the other node in this network. Thanks again, Hosnieh -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
