Slashdot Mirror
Tor: A JAP Replacement
kid_wonder writes "Wired is running an article describing an answer to this previous /. story. Packets are sent through a network of randomly selected servers each of which knows only its predecessor and successor. Packets are unwrapped by a symmetric encryption key at each server that peels off one layer and reveals instructions for the next downstream node. As a 'connection-based low-latency anonymous communication system,' Tor seems to be the answer to JAP to allow anonymous networking activities of all kinds."
That's exactly what I thought (and I believe that we're right). What's interesting here though is that it claims to be low-latency, a quality rarely associated with Freenet and probably the primary reason that Freenet remains largely used by people who need/want _extreme_ anonymity rather than your average movie downloader wanting to avoid one of those nasty lawsuits.
Such systems right now have too high a latency and too much overhead (such as a peer sending "noise" into the network when not having the need to send any real data, just to deter packet analysis) that they aren't terribly practical... for now. So you most likely won't see the technology bundled in the next KaZaA, BitTorrent, etc., but we'll see what the future holds.
- sm
As far as I'm aware Freenet stores encrypted content on each node, not just routing requests through nodes.
Onion routing does just that, it is a method for picking an anonymous route. Freenet is a distributed database.
In onion routing the client picks N nodes from the list of servers and encrypts using each servers public key. Then sends the data to the first server. In onion routing each packet of data contains the entire routing list, though it is encrypted in such a way that each node can only tell what the next node is.
Each Freenet nodes caches data blocks based on demand. When a request arrives looking for a data block Freenet forwards the request to a node that has similar information until the correct block is found. Each freenet node only knows about the next and previous nodes, and the route is determined by the key you are searching for.
09F91102 no, 455FE104 nope, F190A1E8 uh-uh, 7A5F8A09 that's not it, C87294CE no. Ah! 452F6E403CDF10714E41DFAA257D313F.
currently N=3 on tor...
Douglas Calvert
Something named "My own private Idaho", an anonymous remailing software from 1996-1998, did (and is still doing) exactly the same thing, with PGP integration, and server key publication.
I am using tor right now to read slashdot as well as IRC and GAIM. Tor is not supposed to be as low latency as your normal connection. Security is a trade off the slight degradation in latency is worth the improved anonymity...
Douglas Calvert
To quote Paul Syverson from his PET talk:
"The man needs your cover traffic just as much as you need the man for his cover traffic.."
Douglas Calvert
Freenet doesn't use onion routing (last time I checked), but it does use the concept of sending messages through mutiple hops. But the main difference between Freenet and Tor is that Freenet is an anonymous publishing system and Tor is an anonymizing layer that can work with almost any application.
From what I gather from the limited technical details in the article, this protocol would require knowledge of the entire route at the initial node to handle the 'onion layer' encryption.
Correct. The sender wraps the whole onion, and each router removes one layer.
Is there some way of optimizing a path through a given number of nodes without keeping huge amounts of information about latency on every two nodes, or is this just bouncing the packet around for a while for anonymity and accepting the added latency?
It's more like the latter. Optimizing for performance tends to be at odds with anonymity.
http://www.linux-mag.com/2000-04/gear_01.html
I believe the encryption is layered on from the start, and peeling occurs at each transfer, not peel/crypt/peel/crypt/etc.
I was surprised to see no one posted this earlier; the author of Tor gave a very good presentation at DEFCON last week, and I'll have to get out my CD with his presentation on it, but it's different from Freenet in a few ways. For one, apparently Freenet isn't totally free.
As a side-note, the author is still working on a method to accept/sign-up/recruit primary [trusted] nodes.
Linux: The world's best text-adventure game.
Isn't this onion routing thing exactly what freenet uses?
Not in the same form.
Freenet allows posting of data, which does travel through multiple nodes, much like this one. It also allows retrieval of data. However, the two are separate operations. You don't establish a connection between the publisher of data and the reciever, which means Freenet tends to be unsuitable for things that require even remotely interactive latency. I think Tor might wind up being a bit high for, say, SSH, but it could easily be just fine for instant messaging -- two people that don't know each other by anything but pseudonyms and cannot trace each other can conduct conversations.
May we never see th
Actually, freenet hides everything so well, that of course people are going to stash rotten files all over it. An anonymous network doesn't have to be so opaque to the users themselves though, only to eavesdroppers listening from the outside. Take my own idea about a network, metanet. It still guarantees a useful anonymity, but being an IP network (and not some asshat file-trading "p2p is the wave of the future" application), most people on it tend to know everything that is there, without spending 24 hours a day searching hashes.
It's not a climate that encourages kiddy porn. We all have persistent identities (note: to those that think anonymity precludes this, think about it. Anonymity is the inability to link such a nickname to the real person), and no one wants to tarnish them.
And by the way, assuming that my guess is correct and that you're in Canada, you're invited should you ever want to try it out. 20 minute install, and pings to the farthest reaches are 500ms.