Hacker News Reader

Comment by fru2013 | Hacker News Reader

fru2013·

Likely a technical misunderstanding by the author. It is possible for an ISP to detect tor traffic[1], but not determine the contents of the traffic. My guess is that the author is not aware of this.

http://www.netresec.com/?page=Blog&month=2013-04&post=Detect...

View on HN

Considering the entire blog is about 'deep web' you'd hope the author is aware of this difference and made an error. But clicking on his bio it seems to be this authors first post there, so who knows.

> but not determine the contents of the traffic.

Not even this much, actually. Tor doesn't do any sort of protection on the traffic; it just routes it cleverly. The packets themselves are still plaintext if you're not using an encrypted protocol (e.g. HTTPS) on top of Tor.

In practice, most people use Tor through the browser bundle, which includes an HTTPS-Only extension. But if you're routing random other protocols through Tor (e.g. XMPP, BitTorrent, etc.) be aware that Tor is not doing any protection of your plaintext; that's the responsibility of your client/protocol.

Not exactly.

Tor encrypts your traffic within the Tor network. Even if you don't encrypt, Tor does automatically so your ISP cannot see your content.

If you don't encrypt, your content will be visible to the exit node, but not to your ISP.

https://www.torproject.org/about/overview

You're right with respect to the exit node, but here the story is about the entry node. The entry node would not be able to see the traffic contents, because they're encrypted with the exit node's key.

I helped make this diagram to demonstrate these details:

https://www.eff.org/pages/tor-and-https

Reply on news.ycombinator.com