papers/lux-performance-security-tradeoffs/sections/07-triple-proof-overhead.tex

\section{Triple-Proof Overhead}


The complete triple-proof epoch finality mechanism combines BLS + Groth16(ML-DSA + Pulsar)~\cite{lux-benchmarks}:

\begin{center}
\begin{tabular}{lrr}
\toprule
\textbf{Component} & \textbf{Size} & \textbf{Time} \\
\midrule
BLS aggregated verify ($N$=100) & 48 B & 875\,$\mu$s \\
Groth16 verify (3 pairings, BLS12-381) & 192 B & ${\sim}1$--$3$\,ms (CPU) \\
Epoch metadata & $\sim$8 B & --- \\
\midrule
\textbf{Total epoch proof} & \textbf{$\sim$248 B} & \textbf{${\sim}2$--$5$\,ms (CPU)} \\
\bottomrule
\end{tabular}
\end{center}

For comparison, a single Ethereum beacon chain epoch (32 slots, 12s each = 6.4 minutes) requires $\sim$12 KB of aggregate attestation data. The Lux epoch proof (1 second, 248 bytes) is 48$\times$ smaller at 384$\times$ higher frequency.