A Compact and High-Speed TRNG for FPGA Using Grouped Multi-Node Fast Carry Chains
IEEE Access · 2026 · 14 · 100351-100363
DOI: 10.1109/access.2026.3708896 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Kuo-Yu Liao und 3 weiteren Autor:innen. Die Arbeit nennt 41 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Science and Technology Council, Taiwan.
Abstract (Originalsprache)
Many digital services and applications now rely on cryptographic systems. More efficient systems are being developed to meet the growing demand for a better user experience. Therefore, hardware accelerators for cryptographic systems have been introduced and widely explored. For reconfigurability, cryptographic algorithms are often implemented via a field-programmable gate array (FPGA) and embedded in the existing system. True random number generators (TRNGs), as a security fundamental of these cryptographic algorithms, are included in such designs to provide nonces. In this paper, we present a compact, high-speed TRNG architecture based on a fast carry chain that forms a multi-node structure. We transform the fast carry chains into multiple multiplexer-based ring oscillators (ROs) to introduce additional jitter sources and increase randomness. To achieve compact resource utilization and higher throughput, we developed a grouping scheme to manage the outputs of these multi-node ROs. Our design is implemented on the Xilinx Zynq-7000 FPGA, where area performance is measured using configurable logic blocks (CLBs), which explicitly contain 2 SLICEs in Xilinx 7-series products. We collected the random numbers (RNs) using a simple system comprising a processor and a direct memory access (DMA) unit. We evaluated them by conducting the NIST SP 800-22, NIST SP 800-90B, and AIS-31 T8 tests on the RNs. The measurement results indicate that our TRNG design successfully passed the three tests, achieving a minimum bit entropy of 0.9995 and a minimum byte entropy of 8.000. Compared to other works, our TRNG consumes only 12 SLICEs and achieves up to 800 Mbps throughput. In terms of a more universal metric, our TRNG achieves a 4.79x to 9.13x improvement in the power-related figure of merit (FoM).
Förderung
- National Science and Technology Council, Taiwan
Zitieren
APA
Liao, K., Wang, S., Lee, N., & Chen, P. (2026). A Compact and High-Speed TRNG for FPGA Using Grouped Multi-Node Fast Carry Chains. IEEE Access, 14, 100351-100363. https://doi.org/10.1109/access.2026.3708896
Vancouver
Liao K, Wang S, Lee N, Chen P. A Compact and High-Speed TRNG for FPGA Using Grouped Multi-Node Fast Carry Chains. IEEE Access. 2026;14:100351-100363. doi:10.1109/access.2026.3708896
BibTeX
@article{Liao2026,
author = {Liao, Kuo-Yu and Wang, Shao-Huai and Lee, Narn-Yih and Chen, Pei-Yin},
title = {A Compact and High-Speed TRNG for FPGA Using Grouped Multi-Node Fast Carry Chains},
journal = {IEEE Access},
year = {2026},
volume = {14},
pages = {100351--100363},
doi = {10.1109/access.2026.3708896}
}
EUVIMED