Comments on draft-ietf-sipping-3gpp-r5-requirements-00

[Russ Housley <housley@vigilsec.com>]

I have three comments.

  Section 4.24.1.2 points out that IKE and TLS have too many round trips; 
however, it does not offer a  direction for a solution.  Is a lightweight 
security association establishment protocol needed?  If not, can the 
heavyweight security association establishment be used infrequently and 
amoritize the overhead over a long time period?

  Section 4.24.1.3 is a bit misleading; Diffie-Hellman is a public key 
algorithm.

  In section 4.24.3.1, please add a sentence to explain what is meant by 
replay protection.