FPGA Implementation of Metastability-Based True Random Number Generator

Hisashi HATA; Shuichi ICHIKAWA

doi:10.1587/transinf.E95.D.426

FPGA Implementation of Metastability-Based True Random Number Generator

Hisashi HATA, Shuichi ICHIKAWA

Full Text Views

0

Share
Cite this

Summary :

True random number generators (TRNGs) are important as a basis for computer security. Though there are some TRNGs composed of analog circuit, the use of digital circuits is desired for the application of TRNGs to logic LSIs. Some of the digital TRNGs utilize jitter in free-running ring oscillators as a source of entropy, which consume large power. Another type of TRNG exploits the metastability of a latch to generate entropy. Although this kind of TRNG has been mostly implemented with full-custom LSI technology, this study presents an implementation based on common FPGA technology. Our TRNG is comprised of logic gates only, and can be integrated in any kind of logic LSI. The RS latch in our TRNG is implemented as a hard-macro to guarantee the quality of randomness by minimizing the signal skew and load imbalance of internal nodes. To improve the quality and throughput, the output of 64–256 latches are XOR'ed. The derived design was verified on a Xilinx Virtex-4 FPGA (XC4VFX20), and passed NIST statistical test suite without post-processing. Our TRNG with 256 latches occupies 580 slices, while achieving 12.5 Mbps throughput.

Publication: IEICE TRANSACTIONS on Information Vol.E95-D No.2 pp.426-436

Publication Date: 2012/02/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1587/transinf.E95.D.426

Type of Manuscript: Special Section PAPER (Special Section on Reconfigurable Systems)

Category: Application

Cite this

Copy

Hisashi HATA, Shuichi ICHIKAWA, "FPGA Implementation of Metastability-Based True Random Number Generator" in IEICE TRANSACTIONS on Information, vol. E95-D, no. 2, pp. 426-436, February 2012, doi: 10.1587/transinf.E95.D.426.
Abstract: True random number generators (TRNGs) are important as a basis for computer security. Though there are some TRNGs composed of analog circuit, the use of digital circuits is desired for the application of TRNGs to logic LSIs. Some of the digital TRNGs utilize jitter in free-running ring oscillators as a source of entropy, which consume large power. Another type of TRNG exploits the metastability of a latch to generate entropy. Although this kind of TRNG has been mostly implemented with full-custom LSI technology, this study presents an implementation based on common FPGA technology. Our TRNG is comprised of logic gates only, and can be integrated in any kind of logic LSI. The RS latch in our TRNG is implemented as a hard-macro to guarantee the quality of randomness by minimizing the signal skew and load imbalance of internal nodes. To improve the quality and throughput, the output of 64–256 latches are XOR'ed. The derived design was verified on a Xilinx Virtex-4 FPGA (XC4VFX20), and passed NIST statistical test suite without post-processing. Our TRNG with 256 latches occupies 580 slices, while achieving 12.5 Mbps throughput.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.E95.D.426/_p

Copy

@ARTICLE{e95-d_2_426,
author={Hisashi HATA, Shuichi ICHIKAWA, },
journal={IEICE TRANSACTIONS on Information},
title={FPGA Implementation of Metastability-Based True Random Number Generator},
year={2012},
volume={E95-D},
number={2},
pages={426-436},
abstract={True random number generators (TRNGs) are important as a basis for computer security. Though there are some TRNGs composed of analog circuit, the use of digital circuits is desired for the application of TRNGs to logic LSIs. Some of the digital TRNGs utilize jitter in free-running ring oscillators as a source of entropy, which consume large power. Another type of TRNG exploits the metastability of a latch to generate entropy. Although this kind of TRNG has been mostly implemented with full-custom LSI technology, this study presents an implementation based on common FPGA technology. Our TRNG is comprised of logic gates only, and can be integrated in any kind of logic LSI. The RS latch in our TRNG is implemented as a hard-macro to guarantee the quality of randomness by minimizing the signal skew and load imbalance of internal nodes. To improve the quality and throughput, the output of 64–256 latches are XOR'ed. The derived design was verified on a Xilinx Virtex-4 FPGA (XC4VFX20), and passed NIST statistical test suite without post-processing. Our TRNG with 256 latches occupies 580 slices, while achieving 12.5 Mbps throughput.},
keywords={},
doi={10.1587/transinf.E95.D.426},
ISSN={1745-1361},
month={February},}

Copy

TY - JOUR
TI - FPGA Implementation of Metastability-Based True Random Number Generator
T2 - IEICE TRANSACTIONS on Information
SP - 426
EP - 436
AU - Hisashi HATA
AU - Shuichi ICHIKAWA
PY - 2012
DO - 10.1587/transinf.E95.D.426
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E95-D
IS - 2
JA - IEICE TRANSACTIONS on Information
Y1 - February 2012
AB - True random number generators (TRNGs) are important as a basis for computer security. Though there are some TRNGs composed of analog circuit, the use of digital circuits is desired for the application of TRNGs to logic LSIs. Some of the digital TRNGs utilize jitter in free-running ring oscillators as a source of entropy, which consume large power. Another type of TRNG exploits the metastability of a latch to generate entropy. Although this kind of TRNG has been mostly implemented with full-custom LSI technology, this study presents an implementation based on common FPGA technology. Our TRNG is comprised of logic gates only, and can be integrated in any kind of logic LSI. The RS latch in our TRNG is implemented as a hard-macro to guarantee the quality of randomness by minimizing the signal skew and load imbalance of internal nodes. To improve the quality and throughput, the output of 64–256 latches are XOR'ed. The derived design was verified on a Xilinx Virtex-4 FPGA (XC4VFX20), and passed NIST statistical test suite without post-processing. Our TRNG with 256 latches occupies 580 slices, while achieving 12.5 Mbps throughput.
ER -