Some of you may recall me ranting several years ago about the importance of including anti-surveillance features as mandatory aspects of our protocols. I seem to recall getting (mostly) politely laughed at ...
Anyhow, there's an article on the topic in Wired right now that hints at the commercial reasons why service companies MUST build anti-surveillance into their products in order to maintain market viability -- because their competitors are going to do it and draw away the users. The same can apply to our protocol designs.
he point I'd like to make is that not only should our protocols be designed so as to reduce their surveilability, they should also be design to help other protocols do the same. Encryption, random ports, random packet sizes, random interpacket timings, multipath, and peer-to-peer with relay support. Think about it.