Page 9 - 2024F

2 UEC Int’l Mini-Conference No.53

A True Random Number Generator

with Jitter-Sampling by Ring Generator

∗
Tuan-Kiet DANG and Cong-Kha PHAM
Department of Computer and Network Engineering
The University of Electro-Communications, Tokyo, Japan

Keywords: TRNGs, jitter, metastability, ring generator

1 Introduction
True Random Number Generator (TRNG) is an es-
sential primitive for extracting random bits from
a random entropy source. A robust and reliable
TRNG design provides non-deterministic nonces for
cryptographic applications, especially in the secret
key generation process. This paper presents an
accumulating-jitter-based TRNG design that uses a
ring generator to sample high-frequency oscillators
and obfuscate the sampled bit. Fig.1 describes the Figure 2: Entropy cell and its control sequence.
structure of the proposed design.

Figure 1: 16-state ring generator with four oscillator
injection points.

2 Design Methodology

The proposed TRNG comprises two main compo- Figure 3: NIST SP800-22 random test.
nents: primitive cells designed to extract entropy and
a robust ring generator that acts as an obfuscation
mechanism for the extracted entropy. Fig.2 (left) 4 Conclusions
illustrates the structure of an XOR-latch cell, the The proposed TRNG design is compact and
entropy extractor in the proposed design. It can op- lightweight. It is suitable for low-cost, low-power
erate as a free-running oscillator with random jitter systems to provide a quality and robust entropy for
noise. Fig.2 (right) shows the structure of a 16-state secure applications.
ring generator. The ring generator is an optimized References
version of a well-known digital circuit called Linear
Feedback Shift Register (LFSR). With jitter noises [1] Tuan-Kiet Dang, et.al “A True Random Num-
ber Generator on FPGA with Jitter-Sampling
sampled from multiple XL-cells and the obfuscating
property of a ring generator, the TRNG yields high by Ring Generator” 2024 Asian Hardware Ori-
random-quality bits. ented Security and Trust Symposium (Asian-
HOST), pp. 1-6, 2024, doi: 10.1109/Asian-
3 Random tests HOST63913.2024.10838391
Fig.3 shows the results of a common random test
from the National Institute of Standards and Tech-
nology (NIST) SP800-22. The test consists of 15
sub-tests, with the minimum passing threshold score
is p = 0.01.

∗ The author is supported by (SESS) MEXT Scholarship.