On Tue, Jun 12, 2007 at 08:36:30PM +0530, Udhay Shankar N wrote: > You'd still need [pseudo]random delays to foil traffic analysis, no?
I don't think you can do that with more or less realtime communication. It would do quite well for canned voice, or email/IM type of interactive but not realtime communication. If you're concerned about global-adversary type of attacks, you probably should up your paranoia level way up. If I was worried about that, I'd use offline things like one-time pads and dead maildrops.
