Bridging Theory and Practice in Key Exchange Protocols

Key exchange protocols address a crucial problem in security: how to securely distribute cryptographic keys to remote users. Since the seminal paper by Diffie and Hellman in 1976, this subject has been extensively studied for over thirty years.

Yet, designing a secure key exchange protocol is notoriously difficult. Many proposed schemes have been found with security flaws, including those specified in the international standards. Heuristic designs based on ad-hoc arguments rather than rigorous security proofs are commonly seen as bad practice. However, several "provably secure" key exchange protocols also turn out to be insecure.

So far, almost all key exchange protocols in the past have sidestepped an important engineering principle, namely the sixth principle -- i.e., "Do not assume that a message you receive has a particular form unless you can check this" (Anderson & Needham, 1995).

The importance of the sixth principle has been widely acknowledged by the security community for many years, but in reality, key exchange protocol designers have generally abandoned this prudent principle on the grounds of efficiency -- following the sixth principle would require using Zero Knowledge Proof (ZKP), which is considered too computationally expensive. However, discarding ZKP has the serious consequence of degrading the security, as evident by many reported attacks in the past. All these indicate a gap in the field.

In the project, we propose to bridge the gap by combing the sixth principle and the Public Key Juggling (PKJ) technique. The PKJ technique has proved useful in tacking several important security problems in the past. It can be integrated with the sixth principle in a perfect match: the former serves to optimize the protocol efficiency while the latter underpins the protocol robustness.

In the proposed research, we will apply the sixth principle and the juggling technique to design new key exchange protocols that are robust and efficient. We will develop new formal models to capture the sixth principle, which has been largely neglected by existing model specifications. Finally, we will aim to promote robust and efficient key exchange protocols to the international standards. In particular, our J-PAKE key exchange protocol has stood years of cryptanalysis and has been deployed in practical applications. Its standardization is a natural step forward and will benefit the security industry in general.

Besides the academic impact, the project can also have a significant impact on society and industry. The following are the potential beneficiaries.

1. Security industry;

2. End users of security systems;

3. Open source developers.

First, the security industry will likely benefit most from our work. At the moment, there is an acute lack of standard Password Authenticated Key Exchange (PAKE) protocol that is robust, efficient and free. The proposed standardization of the J-PAKE algorithm in this project will benefit the security industry in general.

Second, end users will benefit from more robust key exchange protocols in protecting their secret data. This can be especially evident in life-critical scenarios such as military personnel at the battle field using key exchange protocols to establish the secure communication between each other.

Third, open source developers will benefit from our developing new key exchange protocols and prototypes. In the past, we have contributed to the open source development by providing free prototypes of the invented protocols. In the future research, we plan to do the same.