CROSSING Research Seminar (Guest Talk)

Organizer: Jacqueline Brendel, CROSSING

Abstract

Field Programmable Gate Arrays (FPGAs) have become an increasingly popular choice for implementing hardware-based security systems due to their reconfigurability, high performance, and flexibility. There are numerous benefits to utilizing FPGAs in these systems, such as enabling parallel processing and pipelining capabilities for complex algorithms and cryptographic functions, delivering greater customizability and adaptability by tailoring a solution to meet the specific requirements of individual applications for diverse cryptography algorithms, security levels, data rates, and data formats.
This presentation will showcase how FPGAs can efficiently implement various algorithms, including CRYSTALS-Kyber (a fourth-round candidate in the PQC competition), SHA-3 (the latest SHA algorithm), and speed up federated learning with a defense framework against poisoning attacks. For CRYSTALS-Kyber and SHA-3, the presenter adopted HW(FPGA)/SW codesign, leveraging RISC-V vector extensions to accelerate the algorithms on one hand while optimizing the HW/SW interface on the other. As a result, the system's performance has been significantly enhanced. Additionally, the presenter considers FLAME, a cutting-edge defense against backdoor attacks in federated learning, and implements it on an FPGA, achieving orders of magnitude improvements in the server's performance. Overall, this presentation demonstrates how FPGAs provide an ideal platform for deploying highly efficient and flexible security systems, emphasizing their pivotal role in advancing modern cryptography and data protection to secure the digital landscape effectively.

Speaker Bio

Huimin Li is a Ph.D. student in the Cyber Security research group at the faculty of EEMCS, Delft University of Technology in the Netherlands. She is working under the supervision of Dr. Stjepan Picek. Before that She was a faculty member at Southwest University of Science and Technology, China, a chip design engineer at HiSilicon, Huawei Technologies Co., Ltd., China, and an FPGA design engineer at Shanghai United Imaging Healthcare Co., Ltd., China. She obtained her master’s degree and bachelor’s degree at Northwestern Polytechnical University, China in March 2012 and July 2009, respectively. Her research focuses on deep learning, side-channel attacks, post-quantum cryptography, RISC-V, and FPGA.