Formalizing and Analyzing Sender Invariance

In many network applications and services, agents that share no secure channel in advance may still wish to communicate securely with each other. In such settings, one often settles for achieving security goals weaker than authentication, such as sender invariance. Informally, sender invariance means that all messages that seem to come from the same source actually do, where the source can perhaps only be identified by a pseudonym. This implies, in particular, that the relevant parts of messages cannot be modified by an intruder. In this paper, we provide the first formal definition of sender invariance as well as a stronger security goal that we call strong sender invariance. We show that both kinds of sender invariance are closely related to, and entailed by, weak authentication, the primary difference being that sender invariance is designed for the context where agents can only be identified pseudonymously. In addition to clarifying how sender invariance and authentication are related, this result shows how a broad class of automated tools can be used for the analysis of sender invariance protocols. As a case study, we describe the analysis of two sender invariance protocols using the OFMC back-end of the AVISPA Tool.