Original URL: http://www.theregister.co.uk/2012/08/20/making_onion_networks_more_secure/

Better onion anonymity possible: researcher

Boost to reliability as well

By Richard Chirgwin

Posted in Security, 20th August 2012 00:34 GMT

Onion routing – which in spite of its DARPA genealogy are disliked by national security types – could be made more anonymous, according to a an Iranian researcher now working on a PhD at Concordia University in Montreal.

In this Arxiv-published paper, Ehsan Saboori and co-author Shahriar Mohammadi propose separating the “request” and “response” paths on onion networks to improve user anonymity.

Saboori’s paper notes that over time, a user of a technology such as the TOR project can be identified by analysis of network traffic.

His proposal depends on trusted supernodes (someone at Concordia could have proofread the paper, since it’s mis-spelled “suppernode” throughout) to maintain peer lists of participants, and provide peer information to a requester joining the network.

The requester then randomly chooses two paths for its communications: one path to contact the responder (the host the peer is trying to contact), and a different path for responses.

The idea is that traffic analysis becomes more complex, since each requester on the network is able to periodically – and randomly – select new paths to the responder. This makes it difficult for an intruder to identify the origin path of transferred data.

While the supernodes, if compromised, could reveal participants in the network, the existence of the peers doesn’t identify what routes were chosen by a given requester at a given point in time. Anonymity would, however, depend on scale: if a supernode architecture has too few peers, they would be relatively easy to identify.

Saboori also notes that the scheme needs multiple supernodes so that they don’t become a single point of failure (allowing the onion network to be attacked by a DoS against a supernode). ®