As the question suggests, I wanted to get a sanity-check on this issue... is it the case that the use of a random token exchange during SKID3 authentication itself constitute a prevention of a replay attack, or is there a window of vulnerability in this process (and post-authentication) to this particular attack as well?
Slides 18 through 23 on this presentation provides an excellent overview of SKID3 authentication: http://www.slideshare.net/sriprasanna/authentication-distributed-computing
