Paper 2013/871 Original: eprint.iacr.org/2013/871 ↗

A Unified Security Model of Authenticated Key Exchange with Specific Adversarial Capabilities

Abstract

The most widely accepted models in the security proofs of Authenticated Key Exchange protocols are the Canetti-Krawczyk and extended Canetti-Krawczyk models that admit different adversarial queries with ambiguities and incomparable strength. It is desirable to incorporate specific and powerful adversarial queries into a single unified security model and establish a more practical-oriented security notion. Concerning the security of one-round implicitly authenticated Diffie-Hellman key exchange protocols, we present a unified security model that has many advantages over the previous ones. In the model, a system environment is set up, all of adversarial queries are practically interpreted and definitely characterized through physical environment, and some rigorous rules of secret leakage are also specified. To demonstrate usability of our model, a new protocol based on the OAKE protocol is proposed, which satisfies the presented strong security notion and attains high efficiency. The protocol is proven secure in random oracle model under gap Diffie-Hellman assumption.