A Key Compromise Impersonation attack against Wang's Provably Secure Identity-based Key Agreement Protocol.

Maurizio Adriano Strangio

In a 2005 IACR report, Wang published an efficient identity-based key agreement protocol (IDAK) suitable for resource constrained devices. The author shows that the IDAK key agreement protocol is secure in the Bellare-Rogaway model with random oracles and also provides separate ad-hoc security proofs claiming that the IDAK protocol is not vulnerable to Key Compromise Impersonation attacks and also enjoys Perfect Forward Secrecy (PFS). In this report, we review the security properties of the protocol and point out that it is vulnerable to Unknown Key Share attacks. Although such attacks are often difficult to setup in a real world environment they are nevertheless interesting from a theoretical point of view so we provide a version of the protocol that fixes the problem in a standard way. We also provide a security proof of the IDAK protocol based on the Gap Bilinear Diffie Hellman and random oracle assumptions in the stronger extended Canetti-Krawczyk security model of distributed computing.

Maurizio Adriano Strangio

Hua Guo, Zhoujun Li, Yi Mu +1 more

Shuai Shao, Bin Tian, Xin Wang +1 more

We point out the RYY+ identity-based (ID-based) key agreement protocol is vulnerable to Key-compromise impersonation (KCI) attack and Intermediate results leakage (IRL) attack. Then we give the improved protocol which provides KCI and IRL resistance. The novel protocol is provably secure under the gap Bilinear Diffie-Hellman assumption in the seCK model, which enhances the adversary's queries comparing with previous ID based AKE models. The proposed scheme also has an advantage over most previous schemes on security and efficiency.

Yang Li, Jianming Zhu, Wei Zhang +1 more

We revisit the identity-based (ID-based) key agreement protocol due to Wang et al. The protocol is highly efficient and suitable for practical applications except being vulnerable to Intermediate results leakage (IRL) and Key-compromise impersonation (KCI) attack. We propose an improved protocol following the Full Dual Exponential Challenge Response (FDCR-1) scheme to ensure the signature change every time. We also prove it to be secure in the strengthened extended Canetti -- Krawczyk (seCK) model provides better support for adversary's query than previous. In addition, the improved protocol also has an advantage over most existing protocols on security and efficiency.

M. Choudary Gorantla, Colin Boyd, Juan Manuel González Nieto