mirror of
https://github.com/lux-apps/papers.git
synced 2026-10-30 18:15:21 +00:00
zeekay
abf99d05fc
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.
32 lines
1.1 KiB
TeX
32 lines
1.1 KiB
TeX
\section{Resilience Analysis}
|
|
|
|
\label{sec:resilience}
|
|
%% =====================================================================
|
|
|
|
\begin{table}[h]
|
|
\centering
|
|
\begin{tabular}{p{3.5cm}p{4.5cm}p{4.5cm}}
|
|
\toprule
|
|
\textbf{Failure} & \textbf{Database-First} & \textbf{Chain-First} \\
|
|
\midrule
|
|
Cache/DB corruption & Data loss (DB is truth) & Rebuild from chain
|
|
(chain is truth) \\
|
|
Chain outage & Trades continue, no settlement & Trades continue,
|
|
settlements queued \\
|
|
Indexer crash & N/A (no indexer) & Resume from cursor, no data loss \\
|
|
Split brain & Ambiguous ownership & Chain wins (deterministic) \\
|
|
Admin tampering & Silent modification & Detected by hash chain \\
|
|
Backup divergence & Possible inconsistency & Cache is disposable;
|
|
chain is the backup \\
|
|
\bottomrule
|
|
\end{tabular}
|
|
\caption{Failure modes: database-first vs.\ chain-first.}
|
|
\label{tab:failures}
|
|
\end{table}
|
|
|
|
The chain-first architecture has a strictly better failure model for
|
|
every scenario except chain outage, where both architectures degrade
|
|
similarly (trades continue, settlement is deferred).
|
|
|
|
%% =====================================================================
|