Prof. Jovan Golic
Security Innovation department in Telecom Italia
Prof. Jovan Golic will give two short presentations in a row.
First Presentation: New paradigm for digital true random number generation
The bottom line of secure cryptographic systems are good sources of true random numbers. Purely digital random number generators are composed of logic gates only and are hence cost effective and very suitable for implementations. However, previously proposed solutions, which exploit the phase and frequency jitter in ring oscillators or the metastability of edge-triggered flip-flops, suffer from a very low entropy rate and sensitivity to environmental conditions. It will be first explained why a ring oscillator, namely, an odd number of inverters connected in a ring, is a poor source of true randomness. Then, new solutions, which are called Fibonacci and Galois ring oscillators, will be defined and some basic design criteria introduced. As in ring oscillators, random bits are extracted by sampling through a D-type flip-flop with an independent clock signal, possibly generated by a ring oscillator. It will then be argued why the new oscillators are capable of producing much higher entropy rates than classical ring oscillators. The new oscillators are suitable for both FPGA and ASIC implementations.
Second Presentation: Random masking on logic gate level
Classical implementations of cryptographic algorithms are vulnerable to side-channel attacks such as timing attacks and power analysis attacks. Random masking of cryptographic operations is a technique that may provide resistance to these attacks. For hardware implementations, the concept of computationally secure random masking of logic gates will be introduced. A number of basic techniques for achieving this objective will then be defined and compared with each other.
Short Bio
Jovan Golic is known for active research in cryptology in the last two decades, especially in the area of stream ciphers. He introduced many novel methods for the analysis and design of stream ciphers, including LFSR-based stream ciphers such as nonlinear filter generators, memoryless combiners, and combiners with memory. He also applied these methods to practical ciphers, including A5/1, RC4, Bluetooth, MUX generator, and MUGI. In the recent years, he also contributed to the area of secure hardware implementations of cryptographic algorithms. He is currently working at the Security Innovation department in Telecom Italia, where he is also involved in network security and biometric authentication.