Abstract

The Signal protocol provides end-to-end encryption for billions of users in popular instant messaging applications like WhatsApp, Facebook Messenger, and Google Allo. The protocol relies on an app-specific central server to distribute public keys and relay encrypted messages between the users. Signal prevents passive attacks. However, it is vulnerable to some active attacks due to its reliance on a trusted key server. A malicious key server can distribute fake keys to users to perform man-in-the-middle or impersonation attacks. Signal applications support an authentication ceremony to detect these active attacks. However, this places an undue burden on the users to manually verify each other's public key. Recent studies reveal that the authentication ceremony is time-consuming and confusing, and almost nobody adopts it. Our goal is to explore various approaches for automatically detecting or preventing fake key attacks. We modified a local copy of the Signal server to demonstrate that active attacks are feasible. We then designed three defenses that automatically detect or prevent the attacks. We completed a threat analysis of the defenses and implemented some proof-of-concept prototypes for two of them. We analyze their strengths and weaknesses and outline avenues for future work.

Degree

College and Department

Physical and Mathematical Sciences; Computer Science

Rights

https://lib.byu.edu/about/copyright/