I've had some discussions with Dan and below are my suggestions to resolve his comments. I'll try to update the document next Monday, 5/3, with these changes and the changes to mandatory attributes (http://www.ietf.org/mail-archive/web/emu/current/msg01472.html) depending upon the outcome of any discussion.
Cheers, Joe 1) Remove section 4.1.2 as discussed in http://www.ietf.org/mail-archive/web/emu/current/msg01367.html 2) Add section 3.9 "3.9 Certificate-less Authentication and Generic EAP Method Extension In some cases the peer will not have a way to verify a server certificate and in some cases a server might not have a certificate to verify. Therefore, it is desirable to support certificate-less authentication. An application for this is credential provisioning where the peer and server authenticate each other with a shared password and credentials for subsequent authentication (e.g. a key pair and certificate or a shared key) can be passed inside the tunnel. Another application is to extend existing strong EAP methods with new features such as channel bindings. Great care must be taken when attempting to perform certificate-less authentication. One way of doing it is to establish the tunnel without full server or client verification and inside the tunnel use an EAP method that performs mutual authentication and key derivation. If this technique is used the inner EAP method MUST provide resistance to dictionary attack and a cryptographic binding between the inner method and the tunnel method MUST be established. In addition the cipher suite used to establish the tunnel MUST derive the master key using contribution from both client and server, as in ephemeral Diffie-Hellman cipher suites. The tunnel method MAY allow for certificate-less authentication." _______________________________________________ Emu mailing list Emu@ietf.org https://www.ietf.org/mailman/listinfo/emu
