Research On PUF-assisted Security For Embedded Systems

Recent years, with the continuous development of integrated circuit technology and wide application of fingerprint recognition, the concept of physical inclinable function(PUF) is proposed. It's becoming an important technology in the field of hardware security by the great property such as physical unclonable, lightweight and low power consumption. We deeply research the principle and the current development of PUF. There are some serious threatens in embedded systems, such as IP theft, reverse engineering attacks and related-key attack and so on. This dissertation proposes a PUF-based hardware-assisted techniques to protect the hardware, software, and runtime aspects of the embedded systems. The key idea is to incorporate Physical Unclonable Functions in the hardware design and the software authoring process of the devices. In this dissertation, the focus is mainly on several problems in embedded systems as follows:Firstly, we propose the basic concept and characteristic of physical unclonable functions. The security threatens are focused after comparison with different kinds of PUF. The above researches contributes to ubiquitous security issue in embedded systems.Then we propose a novel approach for hardware IP protection using gate-level obfuscation, which could make design less intelligible in order to weaken the effect of reverse engineering and protect IP at every stage of the hardware design and manufacturing. Based on physical unclonable function, it includes 3 different kinds of gate-level obfuscation, which is efficient to reverse engineering attacks.Second, we proposed an unclonable and secure method for embedded systems with instruction-level PUF. The premise of the security is that the software and the hardware will authenticate each other at a per-instruction basis using Physical Unclonable Functions(PUFs) that are built into the processor. The mechanism can make sure that hardware and software are trusted.Considering with the doubtful reliability of current PUFs, we design a novel digital PUF. Based on the characteristics of LUT in FPGA, we built the model of LUT networks and implement the model in FPGA. With the change of environment(such as temperature and voltage), it can still output reliable and stable data.Based on SRAM PUF, we propose a high-security and high-throughout hardware true random number generator. True random seeds are extracted from the noise on the start-up pattern of SRAM memories. Then, the true random seeds are used as the input for the digital PUF to generate a stream of true random bits with a high throughput. It's useful in some applications which need a large number of random number, such as sensor networks.This dissertation proposes a PUF-based hardware-assisted techniques to enhance the security of hardware, software, and random number generator. The tests show that PUF could improve the security of embedded systems and have a little impact in cost, power consumption and performance. The research is valuable in the industrial and academic area.