Security researchers ask: Does self-destructing data really vanish?

Researchers this week published a paper describing how they broke Vanish, a secure communications system prototype out of the University of Washington that generated lots of buzz when introduced over the summer for its ability to make data self-destruct. But interesting wasn't good enough for researchers at Princeton University, the University of Texas and the University of Michigan, who wondered how well the system could really stand up to attack. I gave the system a whirl back in July and found it to be pretty interesting. Ed Felten from Princeton describes in the Freedom to Tinker blog how he, a fellow researcher at Princeton and peers at the University of Michigan and University of Texas figured out how to beat Vanish.

Such networks, the same kinds used to share music and other files, change over time as computers jump on or off. Their paper is titled "Defeating Vanish with Low-Cost Sybil Attacks Against Large DHTs."  Vanish exploits the churn on peer-to-peer networks by creating a key whenever a Vanish user puts the system to use and then divvying up that key and spreading across the P2P net. As such, portions of the key disappear forever and the original message can't be unencrypted. This led to some interesting technical discussions with the Vanish team about technical details of Vuze and Vanish, and about some alternative designs for Vuze and Vanish that might better resist attacks." Later, Felten ran into an ex-student now at the University of Texas who happened to be investigating Vanish as well, and they wound up collaborating. "The people who designed Vanish are smart and experienced, but they obviously made some kind of mistake in their original work that led them to believe that Vanish was secure - a belief that we now know is incorrect," Felten writes. Felten wrote that after reading about Vanish during the summer "I realized that some of our past thinking about how to extract information from large distributed data structures might be applied to attack Vanish. [S]tudent Scott Wolchok grabbed the project and started doing experiments to see how much information could be extracted from the Vuze DHT [Vuze is the P2P network used by Vanish and DHT is a distributed hash table]. If we could monitor Vuze and continuously record almost all of its contents, then we could build a Wayback Machine for Vuze that would let us decrypt [vanishing data objects] that were supposedly expired, thereby defeating Vanish's security guarantees." Felten goes on to tell an interesting tale about the timing of this realization and the experiments that followed. "We didn't want to ambush the Vanish authors with our break, so we took them aside at the [Usenix Security conference in Montreal in August] and told them about our preliminary results.

The University of Washington researchers investigated the other researchers' findings, updated Vanish and issued a report of their own on the experience.  Among other things, they came up with a way to make breaking Vanish more expensive, Felten writes. We do encourage researchers, however, to analyze it and improve upon it. The University of Washington researchers sum up their latest findings here as well, noting that Vanish does not have to be wedded to Vuze and in fact might be better based on a hybrid system that uses multiple distributed storage systems.  They write: "However, we recommend that at this time, the Vanish prototype only be used for experimental purposes. We strongly believe that realizing Vanish's vision would represent a significant step toward achieving privacy in today's unforgetful age." For more on network research, read our Alpha Doggs Blog. Follow Bob Brown on Twitter.