Name : Louis GOUBIN, Professor
Louis Goubin is a Professor at Versailles St-Quentin-en-Yvelines University and Cryptography & Security Consultant at Gemalto. A former student of the École normale supérieure (Paris), he holds a PhD in Pure Mathematics from Paris XI University (1995) and an Habilitation to supervise research from Paris VII University (2003). He published more than 40 papers, about the design of new asymmetric cryptosystems, cryptanalysis of existing algorithms and protocols, and the protection of software implementations against physical attacks. He has also filled a dozen of patents on practical cryptology and smart cards.
· L. Goubin, J.-M. Masereel, M. Quisquater, “Cryptanalysis of White Box DES Implementations”. In Proceedings of SAC’2007, LNCS, Springer-Verlag, 2007.
· A. Berzati, C. Canovas, L. Goubin, "Perturbating RSA Public Keys: an Improved Attack". In Proceedings of CHES’2008, LNCS, Springer-Verlag, 2008.
Title of Project : Cryptography for the Security of Embedded Systems
The project aims at extending the security notions and attacks to the case of computer programs that are executed within embedded systems. This context tends to become ubiquitous in a world of "ambient intelligence", with more and more miniaturized computing circuits everywhere, including smart cards (for banking, GSM, UMTS, pay-TV, ID-cards or electronic passports applications), RFID tags, personal digital assistants, MP3 readers, etc.
In this project, we propose to study the three main paradigms of cryptology: encryption, electronic signature and authentication, in the context of embedded systems, and to focus on the special case of executable codes.
Concerning the notion of confidentiality for an executable code, a first research direction consists in studying more deeply a recent problematic, which is beyond the classical security models: the necessity of taking the physical nature of computation into account, due to the state of the art of attacks which are based for instance on the statistical analysis of power consumption. In a second direction, the (static) knowledge of the code must not allow the attacker to understand what the program exactly does. Applications of this concept (called "code obfuscation") are particularly important for intellectual property protection for software, by using in particular watermarking solutions of the software. The project aims at improving these solutions, by analyzing their cryptographic strength, and proposing new schemes.
The integrity and authenticity aspects of the code are also being studied, particularly in the case of an embedded system (such as a smart card) which does not contain the program in its ROM, but uses instead the terminals, which it is connected to, as a source of executable instructions. Protection against malicious instructions is thus vital. Our project aims at studying the new induced threats, and thus the corresponding new strategies to thwart these attacks. Many advantages are brought by this kind of architecture: building a new smart card does not require long delays for masking: correcting bugs becomes straightforward (a simple update in the terminals) et does not imply withdrawing smart cards from the market.