International audience

Abstract A distributed key generation (DKG) protocol is a fundamental building block for many threshold cryptosystems. It allows a set of participants to jointly generate a shared secret key without using a trusted party. In 1991, Pedersen proposed the first DKG protocol called Joint‐Feldman DKG. However, it was proved later that this protocol does not guarantee a uniformly random distribution of generated keys. Despite this flaw in the security requirements, Pedersen's DKG protocol has been used for several years as central component to design threshold cryptosystems. Note that most of the solutions proposed in the literature to improve the Pedersen's DKG protocol have several disadvantages. They use private channels, require participants to reveal their secret shares to solve complaints, and use costly computations. This makes them complex and not easy to use in practical situations. In this paper, we present at first an extended version of Joint‐Feldman DKG that ensures a uniform distribution of the generated keys. Then we present a DKG protocol with public channels that use a new strategy to manage complaints without revealing the shares of the secrets held by honest participants and that clearly identifies dishonest participants. We prove that our solution satisfies the security requirements of DKG protocols. Copyright © 2016 John Wiley & Sons, Ltd.

Electronic voting protocols have many advantages over traditional voting but they are complex and subject to many kinds of attacks. Therefore, the use of formal verification methods is crucial to ensure some security properties. We propose to model a recent protocol of remote electronic voting in the applied Pi-calculus. We focalized on some security properties such as fairness which expresses the impossibility of obtaining partial results, eligibility which requires that only legitimate voters can vote, coercion resistance which ensures that no voter may vote under pressure, and verifiability which supposes that anyone can verify the accuracy of the final result. We proved either manually or using the automated verification tool ProVerif that the protocol satisfies these security properties.