Chaum's protocol for detecting man-in-the-middle: Explanation, demonstration, and timing studies for a text-messaging scenario

Abstract

ABSTRACTThis article explains, demonstrates, and evaluates Chaum’s protocol for detecting a man-in-the-middle (MitM) of text-messaging network communications. MitM attacks pose serious risks to many network communications. Networks often mitigate these risks with robust protocols, such as TLS, which assume some type of public-key infrastructure that provides a mechanism for the authenticated exchange of public keys. By contrast, Chaum’s protocol aims to detect a MitM with minimal assumptions and technology, and in particular without assuming the authenticated exchange of public keys. Chaum assumes that the eavesdropper can “sound like” the communicants but that the eavesdropper cannot fabricate sensible conversations.Using an encryption function and one-way function, Chaum’s protocol works in three phases. In Phase I, the communicants exchange their public keys. In Phase II, each communicant generates a random string. The first communicant cryptographically commits to that string, and sends the string to the other communicant after receiving the other’s string. In Phase III, using any of four different “scenarios” the communicants verify that each possesses the same two strings. The protocol forces any MitM to cause the communicants to possess different pairs of strings. The text-messaging scenario is similar to a forced-latency protocol proposed by Wilcox-O’Hearn in 2003.This article implements and experimentally demonstrates the effectiveness of the third scenario, which uses timing to detect a MitM in text-messaging. Even assuming a MitM can send messages without any network latency, the protocol forces the MitM to cause delays noticeable by the communicants. This article is the first to explain, demonstrate, and evaluate Chaum’s protocol, which Chaum described only in an abandoned and nearly inscrutable patent application.