Research topics

• Design of cryptographic primitives
• Cryptographic analysis
• Secure communication
• Lightweight cryptography
• Wireless security
• Number theory and cryptography

 Research projects

In this project we study, analyze, construct stream ciphers and related toolkits. We derive a framework to identify and construct optimal distinguishers between random sources. The notion of sequential distinguisher and the link between distinguishers and hypothesis testing is identified. We provide attacks against the standard Bluetooth encryption E0. We construct a trapdoor stream cipher, that is a new public-key cryptosystem for which encryption is low-cost since entirely based on stream cipher techniques.

In this project we study alternate mechanisms to set up a secure communication channel, that is to establish a symmetric key. We derive and analyze a protocol to authenticate an arbitrary digital information over an insecure channel provided that we have a secure channel at disposal which can be used to authenticate a very short string. This can be used for instance to authenticate a Diffie-Hellman protocol or a public key with (only little) help of human monitoring. Main applications are for local wireless communications and peer-to-peer security. We derive and analyze variants which are non-interactive or with integrated key agreement. We construct series of 3-party protocols including a human operator who can be involved in small computations. We analyze the Bluetooth key establishment protocol and show that security could be higher than expected if properly used.

We study identification protocols over radio link for very low-cost devices. We concentrate on privacy issues and security. We yield flaws in several proposed protocols. We derive a new protocol based on key search with time-memory tradeoffs together with applications and comparisons with alternate protocols. We formalize the notion of privacy and the connection with various communication layers.

In this project we derive and analyze a series of undeniable signature schemes and variants and we study applications. We derive MOVA, a generic undeniable signature scheme with arbitrary small signature length based on a secret group homomorphism which is hard to interpolate. We study several group homomorphisms and implementation issues.

We enjoy breaking cryptographic schemes or studying ways to prove their security. So far, we toyed the Chor-Rivest cryptosystem (one of the earliest cryptosystem based on knapsacks) and TLS (against which we mounted a timing attack over a network by using tools from our optimal distinguisher techniques). We also developed the decorrelation theory.

Besides breaking designs by others, we do commit on cryptographic designs ourselves. Please break us! We designed the block cipher family FOX (now available under the product name IDEA NXT). We designed the DFC block cipher prototype based on decorrelation theory. We designed hash functions based on coding theory.