A Lightweight and High Yield Complementary Metal-Oxide Semiconductor True Random Number Generator with Lightweight...

Area of Science:

• Integrated Circuit Design
• Hardware Security
• Cryptography

Background:

• True Random Number Generators (TRNGs) are crucial for secure cryptographic applications.
• Existing TRNG architectures often face challenges in meeting stringent randomness test requirements.
• Post-processing is essential to enhance the quality of random bits generated by TRNGs.

Purpose of the Study:

• To propose a novel TRNG architecture based on a customized ring oscillator (RO).
• To investigate the effectiveness of a lightweight Photon hash function for post-processing TRNG outputs.
• To evaluate the performance, area, and power consumption of the integrated TRNG design.

Main Methods:

• A customized current-starved ring oscillator (RO) was designed to generate jitter noise.
• A wave converter was employed to derive random outputs from the RO's jitter.
• A lightweight Photon hash function was implemented as a post-processing algorithm for the TRNG.
• The TRNG was fabricated using a 28 nm CMOS process for performance evaluation.

Main Results:

• Initial TRNG samples showed limitations in passing NIST SP 800-22 randomness tests.
• The application of the Photon hash function post-processing significantly improved the pass rate by 50%.
• The integrated design achieved a throughput of 0.0142 Mbps with a power consumption of 31.12 mW and occupied 16,498 µm².

Conclusions:

• The proposed TRNG architecture combined with lightweight Photon post-processing offers a viable solution for generating high-quality random numbers.
• This approach demonstrates a substantial reduction in area (5x) and power consumption (65%) compared to conventional DRBG post-processing.
• The findings highlight the potential of lightweight cryptographic primitives for efficient hardware TRNG design.