John, > Well, I would say that we (HW, SW, Platform providers) cannot expect > to understand all of the ways that their products will be deployed, > so it is extremely hard to state "security is not needed."
That is not what I (and I suspect others) are saying. What I am saying is that security (in practice) turns out to be much harder than "just use IPsec". Really. The corollary to this is that mandating IPsec (at the node level) doesn't actually get you usuable security in IPv6. TO get real security, you have to consider the actual application that needs securing as well as the operational environment where the deployment will take place. There are plenty of applications that already have security that do not use IPsec. Should we/can we force them to use IPsec? No. And if an IPv6 node has limited functionality/purpose, and none of that functionality appears likely to use IPsec (because it has other means for providing security), what is the point of requiring IPsec? I think the big message that people are missing is that IPsec has not become the unbiquitous base-line security that we had once hoped for. And even today, IPv6 only mandates IPsec (with manual keys). No key managment. And if there is one thing we have learned from practical deployments, it's all about key mangement/distribution. That is the hard stuff that makes or breaks usability. Mandating IPsec with just static keying just isn't useful in practice. Thus, continuing to mandate IPsec (while continuing to punt on key management) just looks silly. Thomas -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
