11 dez 2024
08:30 Doctoral defense IC3 Auditorium
Theme
Efficient and Secure Software Implementation of Post-Quantum Cryptosystems
Student
Décio Luiz Gazzoni Filho
Advisor / Teacher
Julio Cesar López Hernandez
Brief summary
"The discovery of a quantum algorithm for efficient solution of the problems of integer factorization and discrete logarithm by P. Shor in the 1990s, combined with advances in the construction of quantum computers, threatens the security of conventional public key cryptosystems based on the presumed difficulty of solving these problems (such as RSA, Diffie-Hellman, DSA and several cryptosystems based on elliptic curves).
In response to this threat, a line of research into cryptosystems capable of resisting attacks by both classical and quantum computers began in the mid-2000s, receiving the name "post-quantum cryptography".
Given the unpredictability of advances in building a cryptographically relevant quantum computer, and the possibility of "harvest now, decrypt later" attacks, the cryptography research community is committed to advancing as quickly as possible in the design, analysis, implementation, and standardization of post-quantum cryptosystems.
However, many of the proposed cryptosystems present efficiency characteristics (such as execution time, memory usage, and size of keys, ciphertexts and signatures) that are partially or totally inferior to the conventional cryptosystems they propose to replace, especially those based on elliptic curves. While there is a decades-long history of research into techniques for implementing the basic operations underlying conventional cryptosystems (arbitrary-precision integer arithmetic), new proposals generally employ new basic operations; from "similar" operations, such as polynomial arithmetic with "small" coefficients (used in several lattice-based cryptosystems), to completely different operations, such as array permutations.
The landscape for cryptographic engineering has changed considerably since the 1970s, when the first conventional public-key cryptosystems were proposed. Research into side-channel attacks, pioneered by P. Kocher in the 1990s demonstrated that attacks on implementations can be the most effective way to violate the security of cryptosystems. Since the 2000s, conventional techniques for improving CPU performance have reached a limit, and it has been necessary to seek new paradigms: vectorization, parallelization, special-purpose instructions, and completely new architectures such as GPUs and matrix multiplication accelerators.
It is in this context that the present research is inserted, proposing new algorithms for calculating fundamental primitives in post-quantum cryptography, as well as software implementation techniques for existing algorithms, on target platforms that range from conventional low-power computing architectures (embedded systems) to matrix multiplication accelerators, proposed for machine learning applications.
Examination Board
Headlines:
| Julio Cesar Lopez Hernandez | IC / UNICAMP |
| Peter Schwabe | MPI-SP/Germany |
| Jeroen Antonius Maria van de Graaf | ICEx/UFMG |
| Marco Aurélio Amaral Henriques | FEEC / UNICAMP |
| Eduardo Moraes de Morais | Inversed Tech |
Substitutes:
| Hilder Vitor Lima Pereira | IC / UNICAMP |
| Marcos Antonio Simplicio Junior | PCS/USP |
| Routo Terada | IME / USP |