mirror of
https://github.com/lux-apps/papers.git
synced 2026-11-02 11:39:52 +00:00
zooqueen
a10bf7bbc6
- canonical lux-corona-pq paper: rank-7 Module-LWE (A in R_q^{8x7}, q=48-bit), Module-SIS
- scrub residual Corona Ring-LWE/R-LWE labels across ethos-of-lux, pqsafe-il5,
lux-whitepaper, lux-quasar-composition, lp-020; Corona<->Pulsar = parameter regime,
both Module-LWE (not rank-1-vs-rank-k)
- LLM.md: drop stale 'Corona as rank-1' wobble note; align to code/Ringtail params
- rebuild affected PDFs
66 lines
2.6 KiB
TeX
66 lines
2.6 KiB
TeX
\section{Empirical Evaluation}
|
|
|
|
\label{sec:evaluation}
|
|
|
|
\subsection{Benchmarks}
|
|
|
|
\begin{table}[h]
|
|
\centering
|
|
\begin{tabular}{@{}lrr@{}}
|
|
\toprule
|
|
\textbf{Operation} & \textbf{Time} & \textbf{Size} \\
|
|
\midrule
|
|
Distributed key generation (3 parties) & 1.2s & --- (no dealer; shares local) \\
|
|
Round 1 (per party, precomputable) & 85ms & commitment $\mathbf{w}_i \in R_q^{M}$ \\
|
|
Round 2 (per party) & 120ms & response $\mathbf{z}_i \in R_q^{N'}$ \\
|
|
Total signing (2-of-3) & 0.6s & --- \\
|
|
Verification & 1.6ms & --- \\
|
|
Signature size & --- & $\sim$33--35\,KB \\
|
|
\bottomrule
|
|
\end{tabular}
|
|
\caption{Corona performance benchmarks (Module-LWE, Category~1; signature size and
|
|
verification time per LP-0120).}
|
|
\label{tab:benchmarks}
|
|
\end{table}
|
|
|
|
\subsection{Geographic Distribution}
|
|
|
|
With 5 parties across US-East, US-West, EU-West, Asia-Pacific, and South America
|
|
(100--250\,ms inter-region latency), end-to-end signing for a 3-of-5 threshold was
|
|
dominated by network round-trips rather than computation, comfortably within the
|
|
finality-certificate cadence. Because Round~1 is message-independent, it can be
|
|
precomputed, leaving only the Round~2 round-trip and the non-interactive combine on the
|
|
critical path.
|
|
|
|
\subsection{Comparison with Classical Threshold Schemes}
|
|
|
|
\begin{table}[h]
|
|
\centering
|
|
\begin{tabular}{@{}lrrrl@{}}
|
|
\toprule
|
|
\textbf{Scheme} & \textbf{Sig Size} & \textbf{Sign Time} & \textbf{Verify} & \textbf{PQ Safe} \\
|
|
\midrule
|
|
BLS Threshold & 48 B & 5ms & 2ms & No \\
|
|
FROST & 64 B & 10ms & 1ms & No \\
|
|
CGGMP21 & 65 B & 200ms & $<$1ms & No \\
|
|
\textbf{Corona} & \textbf{$\sim$33--35\,KB} & \textbf{600ms} & \textbf{1.6ms} & \textbf{Yes} \\
|
|
\bottomrule
|
|
\end{tabular}
|
|
\caption{Threshold signature scheme comparison. Corona is the price of quantum resistance
|
|
on the same dealer-free, two-round profile as FROST.}
|
|
\label{tab:comparison}
|
|
\end{table}
|
|
|
|
Corona's $\sim$33--35\,KB signature is larger than classical schemes by roughly three
|
|
orders of magnitude, but this is acceptable for block certificates (one signature per
|
|
block). Corona's signature is an order of magnitude larger than Pulsar's
|
|
$\sim$3.3\,KB---Pulsar's parameters are tuned for FIPS~204 byte-equality and a compact
|
|
encoding, whereas Corona carries the larger Ringtail-derived module and rejection
|
|
bounds---which is the cost Corona pays for being an independent implementation in the
|
|
Aurora diversity profile. The signing time is dominated by network latency rather than
|
|
computation, and the $\sim$1.6\,ms verification is competitive with classical schemes.
|
|
|
|
%% -----------------------------------------------------------------------
|
|
%% 7. RELATED WORK
|
|
%% -----------------------------------------------------------------------
|