[javajosh]

Is it just me or does this look like a huge opportunity? Last I checked we still have control over our devices, and if they are stupid enough to trust the data they collect, then we should feel free to poison the well. I'm talking about opening random connections to endpoints (either random or those we want to protect), to inject noise into the system. I call the idea "data flak". It could be something as simple as a daemon running in the background, or a browser plugin. You want to spy on my traffic? Fine, good luck picking out my real behavior from the gigabytes of utter crap I'm shoving into your sensors. This works not just at the ISP level, but at every intermediate host, too.

The only counter is for an adversary to own your box, which is far more expensive.

[gruez]

>The only counter is for an adversary to own your box, which is far more expensive.

or require your clients to run your software, like in AOL days

[javajosh]

Well, in general, you'd want to draw a casual link between real physical measurement and network traffic; so yeah, if you own the client (and can accurately determine whether or not it's running in a VM, and/or manipulated by a robot, which is tricky) you can filter out the data flak. If I worked for a data-collection org I'd probably ignore (or blacklist, if I could get away with it) a known source of noise.