mirror of
https://github.com/lux-apps/papers.git
synced 2026-10-30 18:15:21 +00:00
zeekay
4a753bb639
Restructured from flat .tex/.pdf files into per-paper subdirectories. Added Antje Worring as co-author on all papers. Synced PDFs to pdfs/ for download site.
38 lines
1.2 KiB
TeX
38 lines
1.2 KiB
TeX
\section{Empirical Evaluation}
|
|
|
|
\label{sec:evaluation}
|
|
|
|
\subsection{Multi-Valued Consensus Performance}
|
|
|
|
We evaluated Field on a 500-node testnet with $d \in \{2, 4, 8, 16\}$-valued conflicts.
|
|
|
|
\begin{table}[h]
|
|
\centering
|
|
\begin{tabular}{@{}rrrr@{}}
|
|
\toprule
|
|
$d$ & \textbf{Median Rounds} & \textbf{Correctness} & \textbf{TPS} \\
|
|
\midrule
|
|
2 & 8 & 100\% & 2,750 \\
|
|
4 & 10 & 100\% & 2,600 \\
|
|
8 & 12 & 100\% & 2,400 \\
|
|
16 & 15 & 99.99\% & 2,100 \\
|
|
\bottomrule
|
|
\end{tabular}
|
|
\caption{Field protocol performance vs. conflict cardinality $d$}
|
|
\label{tab:multival}
|
|
\end{table}
|
|
|
|
The convergence time grows sub-linearly in $d$, consistent with the $O(\log d)$ factor
|
|
predicted by Theorem~\ref{thm:phase}. All experiments produced the correct outcome
|
|
(highest initial support wins) with $> 99.99\%$ reliability.
|
|
|
|
\subsection{Comparison with Binary Protocol}
|
|
|
|
For $d = 4$, a binary decomposition requires $\log_2 4 = 2$ sequential binary consensus
|
|
instances, each taking $\sim 8$ rounds, totaling $\sim 16$ rounds. The Field protocol
|
|
achieves the same result in 10 rounds, a 37.5\% improvement.
|
|
|
|
%% -----------------------------------------------------------------------
|
|
%% 8. CONCLUSION
|
|
%% -----------------------------------------------------------------------
|