Lets continue the thinking line that assigns to those who think themselves inherently exceptional the inclination, of not the impulse, to now double-down – and get even more devious. Nothing in the character of an exceptional will ever stop, back off and think. Rather, it just plans “more exceptionalism”. In particular, folks retain the triad of exploits: first IPint, then WEBint, then go for HUMint. That is, insert the exceptional exploit into the packet trace. If that doesn’t work, insert the exploit into the many buggy areas of the web. If that doesn’t work, bribe. Of course, be using bribes anyway to engineer a more likely success of the first two by subverting the design process, to engineer-in insecurity.
Next, lets take a hint from my own recent experiences – in that two handsets struggle to develop sound. IN each case, it turned out that minor amounts of physical debris really affect sound deliver (and removal of crud from the “air holes” of the “speaker” made things better. So lets look at the technology – of chip-based microphones and chip-based speakers. Assume intel has microphones built into its ‘chips” that is – as its contribution to American engineered-in insecurity. Since that sounds a little on the paranoid side when stated so broadly, lets take a look at some concrete reality so we don’t sound like a nut (worried about GCHQ microphone implants placed in their recent tooth crowns):
What is also interesting about the paper is its well written summary of how to cryptanalyze streams to distinguish one from another. If you are GCHQ and have implanted an exploit into a device, attempting to subvert the VPN say, on the whole you will be subtly biasing the overall behavior of the processing on the computing device, so that a timing stain is left on such “normal” things as udp packet flows (bearing the data subject to ssl-async security services). A classical covert channel, passive intercepts at the fiber intercept (setup, as a result of targeting) can then recover the timing channel and the keying material – sufficient to allow standard cryptanalysis to break keys efficiently.
Of course, the trick to to keep the covert channel covert!