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.
57 lines
2.2 KiB
TeX
57 lines
2.2 KiB
TeX
\section{Problem: Single-Key Validator Architecture}
|
|
|
|
\label{sec:problem}
|
|
|
|
\subsection{Current Architecture}
|
|
|
|
In a standard PoS validator:
|
|
|
|
\begin{enumerate}[leftmargin=2em]
|
|
\item Key generation: $sk \leftarrow \mathbb{Z}_q$, $pk \leftarrow sk \cdot G$.
|
|
\item Block signing: $\sigma \leftarrow \textsf{Sign}(sk, \text{block})$.
|
|
\item The scalar $sk$ is stored in a single location (keystore file, HSM, or memory).
|
|
\end{enumerate}
|
|
|
|
\subsection{Attack Surface}
|
|
|
|
\begin{table}[ht]
|
|
\centering
|
|
\caption{Validator key compromise attack vectors.}
|
|
\label{tab:attack-vectors}
|
|
\begin{tabular}{lll}
|
|
\toprule
|
|
\textbf{Vector} & \textbf{Mechanism} & \textbf{Impact} \\
|
|
\midrule
|
|
Software exploit & RCE on validator node & Full key extraction \\
|
|
Insider threat & Operator with disk access & Full key extraction \\
|
|
Side channel & Timing/power analysis on signing & Key recovery over time \\
|
|
Cloud provider & VM snapshot, memory dump & Full key extraction \\
|
|
Supply chain & Malicious dependency in node software & Key exfiltration \\
|
|
HSM vulnerability & Firmware exploit (e.g., CVE-2019-15809) & Key extraction \\
|
|
\bottomrule
|
|
\end{tabular}
|
|
\end{table}
|
|
|
|
The common theme: every attack vector leads to \emph{complete} key compromise because
|
|
the key is a single scalar. There is no partial compromise, no graceful degradation.
|
|
|
|
\subsection{Threshold Architecture}
|
|
|
|
In a threshold architecture with parameters $(t, n)$:
|
|
|
|
\begin{enumerate}[leftmargin=2em]
|
|
\item Key generation: Each of $n$ parties receives a share $sk_i$. The master key
|
|
$sk$ is never materialized at any single location.
|
|
\item Signing: Any $t$ parties collaborate to produce a valid signature. The protocol
|
|
output is indistinguishable from a single-signer signature.
|
|
\item Compromise: An adversary who corrupts $< t$ parties learns nothing about $sk$.
|
|
\end{enumerate}
|
|
|
|
The validator's signing capability is distributed across multiple machines,
|
|
geographic locations, or organizational boundaries. Compromising any single
|
|
component yields zero useful information.
|
|
|
|
%% ========================================================================
|
|
%% 3. DISTRIBUTED KEY GENERATION
|
|
%% ========================================================================
|