Spaz Wrote:It depends on what the law would consider to be "peer-to-peer communication". I can only assume that it means communication that goes from Alice to Bob without passing through the provider of some service, because the other points focus on an architecture where there's a middleman. That definition would include things like SSH and HTTPS. I'd like to see some assurance that that isn't the case. And if that definition does NOT include things like SSH and HTTPS, that's a loophole that makes the law essentially useless.
This is what I like about the internet, though. It's easy to work around things like this.
Suppose Alice and Bob want to set up a channel on a peer to peer network on Skype to chat together. Alice sends a message to Bob, but that message gets bounced through Skype. Skype then connects Alice to Eve to Bob. Since Alice and Bob are not directly connected to each other, Alice can be aware of her own internet connections and destroy the call. It's too easy to discover Eve's presence. Alice and Bob can then move to a different network or create their own P2P communication that does not rely on government laws and the possibility of connecting with Eve.
Allowing one connection to be different from another due to these laws would create something different in the communication which allows for detection. A person could write a piece of code notifying Alice that she's actually connected through Eve, then communicate that awareness to Bob. Or Bob could have it as well. It doesn't matter.
This assumes, however, that Alice knows exactly how to route to Bob and is aware of how to protect her own security. This isn't always true.
Also, since you can funnel any kind of communication through SSH tunneling, it becomes hard to detect where a person is truly calling from. You could detect a person through the account name, but accounts are easily disposable.