Berkant Ustaoglu

Obtaining a secure and efficient key agreement protocol from (H)MQV and NAXOS

LaMacchia, Lauter and Mityagin recently presented a strong security definition for authenticated key agreement strengthening the well-known Canetti-Krawczyk definition. They also described a protocol, called NAXOS, that enjoys a simple security proof in the new model. Compared to MQV and HMQV, NAXOS is less efficient and cannot be readily modified to obtain a one-pass protocol. On the other hand MQV does not have a security proof, and the HMQV security proof is extremely complicated. This paper proposes a new authenticated key agreement protocol, called CMQV (‘Combined’ MQV), which incorporates design principles from MQV, HMQV and NAXOS. The new protocol achieves the efficiency of HMQV and admits a natural one-pass variant. Moreover, we present a relatively simple and intuitive proof that CMQV is secure in the LaMacchia-Lauter-Mityagin model.

On the importance of public-key validation in the MQV and HMQV key agreement protocols

HMQV is a hashed variant of the MQV key agreement protocol proposed by Krawczyk at CRYPTO 2005. In this paper, we present some attacks on HMQV and MQV that are successful if public keys are not properly validated. In particular, we present an attack on the two-pass HMQV protocol that does not require knowledge of the victims ephemeral private keys. The attacks illustrate the importance of performing some form of public-key validation in Diffie-Hellman key agreement protocols, and furthermore highlight the dangers of relying on security proofs for discrete-logarithm protocols where a concrete representation for the underlying group is not specified.

Comparing the pre- and post-specified peer models for key agreement

In the pre-specified peer model for key agreement, it is assumed that a party knows the identifier of its intended communicating peer when it commences a protocol run. On the other hand, a party in the post-specified peer model for key agreement does not know the identifier of its communicating peer at the outset, but learns the identifier during the protocol run. In this article, we compare the security assurances provided by the Canetti-Krawczyk security definitions for key agreement in the pre- and post-specified peer models. We give examples of protocols that are secure in one model, but insecure in the other. We also enhance the Canetti-Krawczyk security models and definitions to encompass a class of protocols that are executable and secure in both the pre- and post-specified peer models.

Anonymity and one-way authentication in key exchange protocols

Key establishment is a crucial cryptographic primitive for building secure communication channels between two parties in a network. It has been studied extensively in theory and widely deployed in practice. In the research literature a typical protocol in the public-key setting aims for key secrecy and mutual authentication. However, there are many important practical scenarios where mutual authentication is undesirable, such as in anonymity networks like Tor, or is difficult to achieve due to insufficient public-key infrastructure at the user level, as is the case on the Internet today. In this work we are concerned with the scenario where two parties establish a private shared session key, but only one party authenticates to the other; in fact, the unauthenticated party may wish to have strong anonymity guarantees. We present a desirable set of security, authentication, and anonymity goals for this setting and develop a model which captures these properties. Our approach allows for clients to choose among different levels of authentication. We also describe an attack on a previous protocol of Øverlier and Syverson, and present a new, efficient key exchange protocol that provides one-way authentication and anonymity.

Comparing the Pre- and Post-specified Peer Models for Key Agreement

In the pre-specified peer model for key agreement, it is assumed that a party knows the identifier of its intended communicating peer when it commences a protocol run. On the other hand, a party in the post-specified peer model for key agreement does not know the identifier of its communicating peer at the outset, but learns the identifier during the protocol run. In this paper we compare the security assurances provided by the Canetti-Krawczyk security definitions for key agreement in the pre- and post-specified peer models. We give examples of protocols that are secure in one model but insecure in the other. We also enhance the Canetti-Krawczyk security models and definitions to encompass a class of protocols that are executable and secure in both the pre- and post-specified peer models.

Modeling leakage of ephemeral secrets in tripartite/group key exchange

Recent advances in the design and analysis of secure two-party key exchange (2KE) such as the leakage of ephemeral secrets used during the attacked sessions remained unnoticed by the current models for group key exchange (GKE). Focusing on a special case of GKE-- the tripartite key exchange (3KE) -- that allows for efficient one-round protocols, we demonstrate how to incorporate these advances to the multi-party setting. From this perspective our work closes the most pronounced gap between provably secure 2KE and GKE protocols. The proposed 3KE protocol is an implicitly authenticated protocol with one communication round which remains secure even in the event of ephemeral secret leakage. It also significantly improves upon currently known 3KE protocols, many of which are insecure. An optional key confirmation round can be added to our proposal to achieve the explicitly authenticated protocol variant.

Towards Denial-of-Service-Resilient Key Agreement Protocols

Denial of service resilience is an important practical consideration for key agreement protocols in any hostile environment such as the Internet. There are well-known models that consider the security of key agreement protocols, but denial of service resilience is not considered as part of these models. Many protocols have been argued to be denial-of-service-resilient, only to be subsequently broken or shown ineffective. In this work we propose a formal definition of denial of service resilience, a model for secure authenticated key agreement, and show how security and denial of service resilience can be considered in a common framework, with a particular focus on client puzzles. The model accommodates a variety of techniques for achieving denial of service resilience, and we describe one such technique by exhibiting a denial-of-service-resilient secure authenticated key agreement protocol. Our approach addresses the correct integration of denial of service countermeasures with the key agreement protocol to prevent hijacking attacks that would otherwise render the countermeasures irrelevant.

Multi-party off-the-record messaging

Most cryptographic algorithms provide a means for secret and authentic communication. However, under many circumstances, the ability to repudiate messages or deny a conversation is no less important than secrecy and authenticity. For whistleblowers, informants, political dissidents and journalists --- to name a few --- it is most important to have means for deniable conversation, where electronic communication must mimic face-to-face private meetings. Off-the-Record Messaging, proposed in 2004 by Borisov, Goldberg and Brewer, and its subsequent improvements, simulate private two-party meetings. Despite some attempts, the multi-party scenario remains unresolved. In this paper, we first identify the properties of multi-party private meetings. We illustrate the differences not only between the physical and electronic medium but also between two- and multi-party scenarios, which have important implications for the design of private chatrooms. We then propose a solution to multi-party off-the-record instant messaging that satisfies the above properties. Our solution is also composable with extensions that provide other properties, such as anonymity.

Security arguments for the UM key agreement protocol in the NIST SP 800-56A standard

The Unified Model (UM) key agreement protocol is an efficient Diffie-Hellman scheme that has been included in many cryptographic standards, most recently in the NIST SP 800-56A standard. The UM protocol is believed to possess all important security attributes including key authentication and secrecy, resistance to unknown key-share attacks, forward secrecy, resistance to known-session key attacks, and resistance to leakage of ephemeral private keys, but is known to succumb to key-compromise impersonation attacks. In this paper we present a strengthening of the Canetti-Krawczyk security definition for key agreement that captures resistance to all important attacks that have been identified in the literature with the exception of key-compromise impersonation attacks. We then present a reductionist security proof that the UM protocol satisfies this new definition in the random oracle model under the Gap Diffie-Hellman assumption.

On reusing ephemeral keys in Diffie-Hellman key agreement protocols

A party may choose to reuse ephemeral public keys in a Diffie-Hellman key agreement protocol in order to reduce its computational workload or to mitigate against denial-of-service attacks. In this note, we show that small-subgroup attacks can be successfully launched on some Diffie-Hellman protocols that reuse ephemeral keys if domain parameters are not appropriately selected or if public keys are not appropriately validated.