The problem with this idea is that as TPB migrates to additional domains, it leaves open a few possibilities. First, sites opposed to TPB will create malicious sites to try and spoof TPD to collect data, spread malware, and otherwise degrade the services that TPB offers. Second, while rapid DNS techniques are well implemented in malware like botnets, it's not a capability that the average Joe can keep track of, so without some form of front end that can track and change with the rotating DNS entries, people will get frustrated and stop using the service. And this is exactly what the opposition wants. It's time that many of these types of technologies go through redesign again. And they must be redesigned from the protocol level. TPB succeeded where Napster didn't because the content being traded was not hosted with TPB. But they are still centrally located for the purpose of search and front end to the users. Anytime a bellybutton can be poked, it will be. So the protocol must be designed to prevent ANY centralized management. Next, any protection mechanisms (trust models, etc) that are built into the protocols used must be extremely well thought out. When analyzing for protocol weakness, rules matter. If only certain nodes are allowed to be "trusted" ,then you can better bet that those wishing to exploit it will design a node to be trusted. Protocols must also be non-differential at all levels, from the handshake of the SSL layer down to the initial HELO to the transmission characteristics of how it sends data. If the protocol can be differentiated, it can be blocked, tracked, hacked, or otherwise interfered with. Simply wrapping it in Tor isn't going to work. There are ample talks out there on how Tor traffic can be characterized and interfered with.