Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

CROSS PQC: An FPGA Implementation of the Message Signing Algorithm

Mortey, Jedidiah Buernor LU and Cene, Xhulian LU (2026) EITM02 20261
Department of Electrical and Information Technology
Abstract
Continuous advancement in Quantum Computing research has exposed vulnera-
bilities within the security architecture of critical systems on which the day-to-day
operations rely. This has set a challenge for the security community to design
and implement alternatives that mitigate the risks associated with the eventual
development of cryptographically significant quantum computers. Hence, NIST
initiated the process to obtain new security standards resistant to the threat vec-
tors enabled by quantum computing and has since standardized several solutions
and are currently in the process of standardizing fallback solutions should, in case,
the first set of standards are compromised as a majority are based on similar
concepts.
... (More)
Continuous advancement in Quantum Computing research has exposed vulnera-
bilities within the security architecture of critical systems on which the day-to-day
operations rely. This has set a challenge for the security community to design
and implement alternatives that mitigate the risks associated with the eventual
development of cryptographically significant quantum computers. Hence, NIST
initiated the process to obtain new security standards resistant to the threat vec-
tors enabled by quantum computing and has since standardized several solutions
and are currently in the process of standardizing fallback solutions should, in case,
the first set of standards are compromised as a majority are based on similar
concepts.
This thesis investigates one of the newer candidates for standardization, the
Code and Restricted Objects Signature Scheme, and attempts to implement the
message signing algorithm of the scheme to achieve a higher performance relative
to existing implementations.
We analyzed the algorithm and implemented incrementally for all defined se-
curity levels in the fast mode of the algorithm (RSDP-fast). We then implemented
the design and ran it on Xilinx ZCU102 FPGA evaluation board. We also verified
its correctness by comparing the results to available Known-Answer Tests. The
implemented design, averagely, used 140% more LUTs, 150% more Flip flops, but
27% less BRAM tiles more than the reference design however took averagely 14%
more time at levels 1 and 5, and 0.1% less time at level 3 to execute than the fastest
existing implementation. The design implemented in this project was synthesized
to run at 125MHz. (Less)
Popular Abstract
The security of critical systems is founded on various mathematical and compu-
tational assumptions which have, in recent years, faced the threat of failure due
to developments in quantum computing. Researchers have shown that security
schemes like RSA as well as schemes based on elliptic curves, which are in use
today, can be broken if large enough quantum computers can be built. Hence, the
security community has been faced with a challenge to implement resolutions to
the risk posed by quantum computers. The National Institute of Standards and
Technology (NIST) has so far led efforts in standardizing proposed solutions to
replace the endangered cryptosystems. The standardized solutions have largely
been based on lattice... (More)
The security of critical systems is founded on various mathematical and compu-
tational assumptions which have, in recent years, faced the threat of failure due
to developments in quantum computing. Researchers have shown that security
schemes like RSA as well as schemes based on elliptic curves, which are in use
today, can be broken if large enough quantum computers can be built. Hence, the
security community has been faced with a challenge to implement resolutions to
the risk posed by quantum computers. The National Institute of Standards and
Technology (NIST) has so far led efforts in standardizing proposed solutions to
replace the endangered cryptosystems. The standardized solutions have largely
been based on lattice structures and stateless hash algorithms. NIST, in a bid to
ensure the availability of fallback solutions if the security of the earlier standard-
ized solutions is to be compromised. One of the solutions submitted in response to
the second request is known as the Code and Restricted Object Signature Scheme
(CROSS).
A signature scheme provides a means by which a sender can encode a proof
of origin for a message, and a receiver can verify that a message truly originates
from the sender. The way CROSS does this is by providing three algorithms: a
key generator for creating a secret key belonging to the sender and a public key
used by any valid receiver, a message signing algorithm which enables the sender
to generate a signature using the secret key and any message such that a receiver
can use the public key to verify the provided signature or signed message.
Algorithms can generally be executed at some speed by general purpose pro-
cessors like those found in everyday computers. However, specialized processors,
also called accelerators, can often be developed to greatly improve the usage of
said algorithms. This project sought to implement the message signing algorithm
of the scheme in hardware, targeting an increase in performance. During test-
ing, we found that the main bottleneck in the implemented design as compared
to the fastest implementation provided by Karl et al. was the width of the data
applied to matrix operations. This points towards a viable area for optimization
in subsequent implementations. (Less)
Please use this url to cite or link to this publication:
author
Mortey, Jedidiah Buernor LU and Cene, Xhulian LU
supervisor
organization
course
EITM02 20261
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Code and Restricted Objects Signature Scheme, CROSS, Digital Signature Algorithm, DSA, Post-Quantum Cryptography, PQC, FPGA
report number
LU/LTH-EIT 2026-1191
language
English
id
9251173
date added to LUP
2026-09-17 10:24:14
date last changed
2026-09-17 10:24:14
@misc{9251173,
  abstract     = {{Continuous advancement in Quantum Computing research has exposed vulnera-
bilities within the security architecture of critical systems on which the day-to-day 
operations rely. This has set a challenge for the security community to design 
and implement alternatives that mitigate the risks associated with the eventual 
development of cryptographically significant quantum computers. Hence, NIST 
initiated the process to obtain new security standards resistant to the threat vec-
tors enabled by quantum computing and has since standardized several solutions 
and are currently in the process of standardizing fallback solutions should, in case, 
the first set of standards are compromised as a majority are based on similar 
concepts.
This thesis investigates one of the newer candidates for standardization, the 
Code and Restricted Objects Signature Scheme, and attempts to implement the 
message signing algorithm of the scheme to achieve a higher performance relative 
to existing implementations.
We analyzed the algorithm and implemented incrementally for all defined se-
curity levels in the fast mode of the algorithm (RSDP-fast). We then implemented 
the design and ran it on Xilinx ZCU102 FPGA evaluation board. We also verified 
its correctness by comparing the results to available Known-Answer Tests. The 
implemented design, averagely, used 140% more LUTs, 150% more Flip flops, but 
27% less BRAM tiles more than the reference design however took averagely 14% 
more time at levels 1 and 5, and 0.1% less time at level 3 to execute than the fastest 
existing implementation. The design implemented in this project was synthesized 
to run at 125MHz.}},
  author       = {{Mortey, Jedidiah Buernor and Cene, Xhulian}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{CROSS PQC: An FPGA Implementation of the Message Signing Algorithm}},
  year         = {{2026}},
}