Deriving Ephemeral Authentication Using Channel Axioms

As computing and computer networks become more and more intertwined with our daily lives, the need to develop flexible and on-the-fly methods for authenticating people and their devices to each other has become increasingly pressing. Traditional methods for providing authentication have relied on very weak assumptions about communication channels, and very strong assumptions about secrecy and the availability of trusted authorities. The resulting protocols rely on infrastructures such as shared secrets and public key hierarchies that are too rigid to support the type of flexible ad-hoc communication we are growing accustomed to and beginning to rely upon.

Recently, different families of protocols allow us to weaken assumptions about trusted infrastructure by strengthening the assumptions about communication channels. Examples include proximity verification protocols, that rely, for example, on the round trip time of a challenge and response; and bootstrapping protocols that rely upon human-verifiable channels, that is, low-bandwidth communication between humans. The problem now becomes:

How do we ensure that the protocols are achieve their security goals?

A vast amount of literature exists on the formal analysis of cryptographic protocols, and mathematical foundations of protocol correctness, but almost all of it relies upon the standard assumptions about the channels in end-to-end, and so its usefulness for nonstandard channels in pervasive networks is limited. In this paper, we present some initial results of an effort towards a formalizing the reasoning about the security of protocols over nonstandard channels.