Ian Martiny (University of Colorado Boulder), Gabriel Kaptchuk (Boston University), Adam Aviv (The George Washington University), Dan Roche (U.S. Naval Avademy), Eric Wustrow (University of Colorado Boulder)

The Signal messaging service recently deployed a emph{sealed sender} feature that provides sender anonymity by cryptographically hiding a message's sender from the service provider. We demonstrate, both theoretically and empirically, that this one-sided anonymity is broken when two parties send multiple messages back and forth; that is, the promise of sealed sender does not emph{compose} over a conversation of messages. Our attack is in the family of Statistical Disclosure Attacks (SDAs), and is made particularly effective by emph{delivery receipts} that inform the sender that a message has been successfully delivered, which are enabled by default on Signal. We show using theoretical and simulation-based models that Signal could link sealed sender users in as few as 5 messages.

Our attack goes beyond tracking users via network-level identifiers by working at the application layer of Signal. This make our attacks particularly effective against users that employ Tor or VPNs as anonymity protections, who would otherwise be secure against network tracing. We present a range of practical mitigation strategies that could be employed to prevent such attacks, and we prove our protocols secure using a new simulation-based security definition for one-sided anonymity over any sequence of messages. The simplest provably-secure solution uses many of the same mechanisms already employed by the (flawed) sealed-sender protocol used by Signal, which means it could be deployed with relatively small overhead costs; we estimate that the extra cryptographic cost of running our most sophisticated solution in a system with millions of users would be less than $40 per month.

View More Papers

Is Your Firmware Real or Re-Hosted? A case study...

Abraham A. Clements, Logan Carpenter, William A. Moeglein (Sandia National Laboratories), Christopher Wright (Purdue University)

Read More

Data Poisoning Attacks to Deep Learning Based Recommender Systems

Hai Huang (Tsinghua University), Jiaming Mu (Tsinghua University), Neil Zhenqiang Gong (Duke University), Qi Li (Tsinghua University), Bin Liu (West Virginia University), Mingwei Xu (Tsinghua University)

Read More

Towards Measuring Supply Chain Attacks on Package Managers for...

Ruian Duan (Georgia Institute of Technology), Omar Alrawi (Georgia Institute of Technology), Ranjita Pai Kasturi (Georgia Institute of Technology), Ryan Elder (Georgia Institute of Technology), Brendan Saltaformaggio (Georgia Institute of Technology), Wenke Lee (Georgia Institute of Technology)

Read More