International audienceIn the domain of smart cards, secured devices must be protected against high level attack potential [1]. According to norms such as the Common Criteria [2], the vulnerability analysis must cover the current state-of-the-art in term of attacks. Nowadays, a very classical type of attack is fault injection, conducted by means of laser based techniques. We propose a global approach, called Lazart, to evaluate code robustness against fault injections targeting control flow modifications. The originality of Lazart is twofolds. First, we encompass the evaluation process as a whole: starting from a fault model, we produce (or establish the absence of) attacks, taking into consideration software countermeasures. Furthermore, according to the near state-of-the-art, our methodology takes into account multiple transient fault injections and their combinatory. The proposed approach is supported by an effective tool suite based on the LLVM format [3] and the KLEE symbolic test generator [4]

International audiencePhysical fault injections break security functionalities of algorithms by targeting their implementations. Software techniques strengthen such implementations to enhance their robustness against fault attacks. Exhaustively testing physical fault injections is time consuming and requires complex platforms. Simulation solutions are developed for this specific purpose. We chose two independent tools presented in 2014, the Laser Attack Robustness (Lazart) and the Embedded Fault Simulator (EFS) in order to evaluate software implementations against multiple fault injection attacks. Lazart and the EFS share the common goal that consists in detecting vulnerabilities in the code. However, they operate with different techniques , fault models and abstraction levels. This paper aims at exhibiting specific advantages of both approaches and proposes a combining scheme that emphasizes their complementary nature

International audienceIndustrial systems are publicly the target of cyberattacks since Stuxnet [1]. Nowadays they are increasingly communicating over insecure media such as Internet. Due to their interaction with the real world, it is crucial to prove the security of their protocols. In this paper, we formally study the security of one of the most used industrial protocols: OPC-UA. Using ProVerif, a well known cryptographic protocol verification tool, we are able to check secrecy and authentication properties. We find several attacks on the protocols and provide countermeasures

International audienceThere exist several automatic verification tools of cryptographic protocols, but only few of them are able to check protocols in presence of algebraic properties. Most of these tools are dealing either with Exclusive-Or (XOR) and exponentiation properties, so-called Diffie-Hellman (DH). In the last few years, the number of these tools increased and some existing tools have been updated. Our aim is to compare their performances by analysing a selection of cryptographic protocols using XOR and DH. We compare execution time and memory consumption for different versions of the following tools OFMC, CL-Atse, Scyther, Tamarin, TA4SP, and extensions of ProVerif (XOR-ProVerif and DH-ProVerif). Our evaluation shows that in most of the cases the new versions of the tools are faster but consume more memory. We also show how the new tools: Tamarin, Scyther and TA4SP, can be compared to previous ones. We also discover and understand for the protocol IKEv2-DS a difference of modelling by the authors of different tools, which leads to different security results. Finally, for Exclusive-Or and Diffie-Hellman properties, we construct two families of protocols P xori and P dhi that allow us to clearly see for the first time the impact of the number of operators and variables in the tools' performances