> > There is a part of me that wishes that we just deploy what we have. There is > another part that notices that between IEEE802, privacy addresses, CGA, stable > privacy addresses and probably a few more, we have built a nice mess and > maybe we should clean it up. For example, we could decree that composing > and IID from a MAC address is obsolete, and we could unify CGA and other > privacy addresses. Add some salt can make the CGA addresses as randomized > as the privacy addresses; change the lifetime of the key and you get whatever > address lifetime is deemed appropriate. But then, the best part of me observes > that with the current level of IPv6 deployment, maybe we don't have an urgent > need for more specs... >
This has been already stated in CGA RFC http://tools.ietf.org/html/rfc3972#section-7.3 . The problem with CGA deployment is not privacy, but the computational costs and the RPKI. It does not need to merge to any other draft to fulfill the privacy. This is why I am trying to address that problem in my other draft and find the simplest way. I think the use of ECC is a good option but I will publish my results soon. Regards, Hosnieh -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------