A Fast Side-Channel Leakage Simulation Technique Based on IC Chip Power Modeling

Hardware-based secure integrated circuit (IC) chips with implemented cryptographic algorithms are vulnerable to practical attacks analyzing side-channel (SC) leakage such as electromagnetic radiation. Correlation power analysis (CPA) is a real threat to secret key crypto ICs. Therefore, the information leakage risks need to be evaluated at the design stage of secure devices. This letter proposes a fast power leakage simulation method for hardware-implemented cryptographic ICs. The power delivery network (PDN) of cryptographic hardware including a silicon substrate is modeled by a chip power model (CPM) and a chip package system (CPS) board model. The proposed method was applied to the advanced encryption standard (AES) test chip implemented by the flip-chip ball grid array (FC-BGA) assembly technology. The simulation results are evaluated by CPA for the SC leakage during plaintext encryption propagating through PDN and silicon substrate. The measurement and simulation successfully find 128-bit secret key on CPA attack. The proposed simulation technique reduces the time required in the exploration of SC leakage risks for secure device designs.

I. Introduction

Hardware security becomes more important than ever in the Internet of Things (IoT) era. The design turnaround time of security oriented devices is shortened before going to manufacturing, under the increasing demands of information devices for mobile communications, connected cars, wearable devices and so forth, all with the higher level of security. Hardware security is hindered in a latent problem at the design stage and may become apparent after the production. Cryptographic algorithms are mathematically secure while vulnerable to hardware attacks, once they are actualized with physical implementation. Side channel (SC) attacks [1] are well known, which an adversary uses to unexpectedly decipher the secret information stored in a cryptographic device. Ubiquitous data processing entities including IoT devices are exposed to the real threat of SC attacks. The security risk and its evaluation methods for those encryption devices have been reported in several studies [2]–[4].