Statement Regarding Temporary Hiatus on U.S. Dollar Withdrawals • Hacker News Board

Re: Statement Regarding Temporary Hiatus on U.S. Dollar Withdrawals

#111

Post by m-r-a-m » Fri, Jun 21, 2013, 7:55 PM UTC

betterunix wrote: post #85

m-r-a-m wrote: post #14

Earlier quoted context omitted.

I don't think the entire usefulness of Bitcoin is related to staying anonymous as it's actually pretty hard to stay anonymous when using Bitcoin. Also, wouldn't the government need to be operating a certain percentage of the Tor nodes to do effective traffic analysis?

"Also, wouldn't the government need to be operating a certain percentage of the Tor nodes to do effective traffic analysis?" Not if they can conduct surveillance on the entire Internet.

You establish a random circuit through the Tor network when you first connect. Even if the NSA is analyzing every packet between every node, they're not going to be able to distinguish your packet from anyone else's. I think they would need to control some of the nodes to keep track of packet sources and destinations.

Re: Statement Regarding Temporary Hiatus on U.S. Dollar Withdrawals

#112

Post by betterunix » Fri, Jun 21, 2013, 11:40 PM UTC

m-r-a-m wrote: post #111

betterunix wrote: post #85

Earlier quoted context omitted.

"Also, wouldn't the government need to be operating a certain percentage of the Tor nodes to do effective traffic analysis?" Not if they can conduct surveillance on the entire Internet.

You establish a random circuit through the Tor network when you first connect. Even if the NSA is analyzing every packet between every node, they're not going to be able to distinguish your packet from anyone else's. I think they would need to control some of the nodes to keep track of packet sources and destinations.

You are wrong. Tor does not include any latency, so a passive global adversary who can watch every packet can correlate your input to the Tor network with the traffic between nodes and the outputs of the exit nodes, and thus break the anonymity property. The reason this does not apply to anonymous remailers is that they randomly permute the messages they receive, sending them out in a different order than they were received; this creates far too much latency for use with HTTP or IRC or other common protocols, so Tor sacrifices security against a global adversary.

http://www.cs.usfca.edu/~ejung/courses/f11683/lectures/tor.p...

http://mice.cs.columbia.edu/getTechreport.php?techreportID=5...

Re: Statement Regarding Temporary Hiatus on U.S. Dollar Withdrawals

#113

Post by sneak » Sat, Jun 22, 2013, 3:01 AM UTC

Robin_Message wrote: post #107

tlrobinson wrote: post #100

Earlier quoted context omitted.

By "forge" I meant using the private keys (obtained through FISA, or possibly cracked) to either decrypt or MITM SSL sessions, not using one of their own CAs to create a valid but obviously fake certificate, which I assume is what you thought I meant. Maybe "forge" was the wrong word.

I think (not an expert) that with elliptic curve Diffie-Hellman (which e.g. Google uses), even with the private key and the whole session, you still can't decrypt the session (Diffie-Hellman provides a secure channel, public/private proves the server is who it says it is).

That is correct. These are referred to as ECDHE, or DHE for the non-ECC variant (vanilla DH). The last e is "ephemeral", referring to the lifetime of the session key.

The property that this provides is known as "forward secrecy".