A method of implanting combinational hardware Trojan based on evolvable hardware

Highlights

• Compared with the general method, it is not to add hardware Trojan, but to modify the original circuit to generate Trojan function.

• The hardware Trojan designed by EHW has little effect on the number of the logic gates and power consumption of the original circuit, and has good concealment.

• According to the symmetric value and similar value to evaluate the difficulty of evolution, choose the hardware Trojan that is easy to evolve.

Abstract

In order to reduce the impact of Hardware Trojan (HT) implantation on the original circuit power consumption, a method using Evolvable Hardware (EHW) to implant combinational HT is proposed. First, the truth table of the original circuit is modified according to the category of HT. Secondly, the symmetric value and similar value are used to evaluate the difficulty degree of evolution, and the truth table which is easy to evolve is selected for evolution. Finally, from the various circuits generated by evolution, the circuit close to the number of logic gates and power consumption of the original circuit is selected to replace the original circuit. The obtained circuit not only includes the HT function, but also improve the anti-power detection ability. According to the experimental results, the HT implanted by EHW has smaller power consumption difference than the HT implanted by general method, and has stronger anti-power detection ability.

Introduction

With the rapid development of electronic technology, integrated circuit (IC) is widely used in every aspects of life, and its security is very important. However, the production process of IC is not safe, which may be implanted into Hardware Trojan (HT), resulting in abnormal circuit function [1], [2], [3], seriously affecting people's normal production and life. Therefore, in recent years, the HT detection has gradually become a research hotspot, such as side-channel analysis (power consumption analysis [4], [5], [6], electromagnetic analysis [7,8],), optical analysis [9,10], logic testing [11], [12], [13]. But these technologies are still in the initial stage and are not perfect. Mainly because real HT exists the situation of sampling difficulty, less varied categories and so on. In order to provide research objects for detection technology, many people have designed HT:

In 2016, Yang et al. [14] constructed a circuit, using a capacitor to absorb the charge during digital conversion in the wire. When the capacitor is fully charged, an attack is launched, forcing the victim's trigger to change to the value the attacker needs.

In 2017, Sepulveda et al. [15] designed an improved "prime + probe" cache attack to attack the 128-bit Advanced Encryption Standard (AES). This design uses bus communication for the first time to improve its efficiency.

In 2019, Yang et al. [16] designed an ordinary HT, which can attack the key based on bit string technology with only one trigger. The purpose is to warn the password designer that it needs to consider the hardware security when designing passwords.

From the above several HT design methods, we can see that they are basically based on functional design, supplemented by hidden design. There is a common problem: additional HT circuit causes power consumption change and anti-detection ability becoming not high. In order to solve this problem, this paper proposes a method of implanting the combinational HT [17] with EHW technology. The original circuit is redesigned to increase the HT function, while keeping the changes of logic gate number and power consumption as small as possible, so as to improve the ability of anti-power detection and provide more hidden HT for detection technology.

The rest of the paper is organized as follows. In Section 2 we analyse the general method, power consumption and EHW technology. In Section 3 we design the steps of implanting HT with EHW technology. In Section 4, the strategy of modifying the truth table is proposed and verified in Section 5. Section 6 verifies the effectiveness of the proposed method by using experiments. Section 7 summarizes the full text and points out the next research direction.