On the Security and Usability of Crypto Phones

Abstract

Crypto Phones represent an important approach for end-to-end VoIP security, claiming to prevent "wiretapping" and session hijacking attacks without relying upon third parties. In order to establish a secure session, Crypto Phones rely upon end users to perform two tasks: (1) checksum comparison: verbally communicating and matching short checksums displayed on users' devices, and (2) speaker verification: ascertaining that the voice announcing the checksum is the voice of the legitimate user at the other end. However, the human errors in executing these tasks may adversely affect the security and usability of Crypto Phones. Particularly, failure to detect mismatching checksums or imitated voices would result in a compromise of Crypto Phones session communications. We present a human factors study, with 128 online participants, investigating the security and usability of Crypto Phones with respect to both checksum comparison and speaker verification. To mimic a realistic VoIP scenario, we conducted our study using the WebRTC platform where each participant made a call to our IVR server via a browser, and was presented with several challenges having matching and mismatching checksums, spoken in the legitimate user's voice, different speakers' voices and automatically synthesized voices. Our results show that Crypto Phones offer a weak level of security (significantly weaker than that guaranteed by the underlying protocols), and their usability is low (although might still be acceptable). Quantitatively, the overall average likelihood of failing to detect an attack session was about 25-50%, while the average likelihood of accepting a legitimate session was about 75%. Moreover, while the theory promises an exponential increase in security with increase in checksum size, we found a degradation in security when moving from 2-word checksum to 4-word checksum.