p2p wiki web

I have been looking into the webrtc and the p2p communication protocol, and I am interested in the idea/possibility of having web content served from individual computers instead of from centralised servers. So for a site like Wikipedia it could be run(loaded/served) completely from/by users, and for a social networking site like facebook posts could be send to just the users you want to send them to. I think you could also have partially peer hosted web pages and use xpath to maps to the nodes on the page you want to replace with user generated content.

Does anyone think something like this could be possible, or have you heard of anything like this already running on the web?

Do you think this sort of network topology for the web would make tracking and storing of web traffic and messages more difficult for governments or large companies?