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.
68 lines
2.9 KiB
TeX
68 lines
2.9 KiB
TeX
\section{Adversary Classes}
|
|
|
|
\label{sec:adversaries}
|
|
|
|
\subsection{Classical Adversary}
|
|
|
|
\begin{definition}[Classical adversary $\mathcal{A}_C$]
|
|
A probabilistic polynomial-time (PPT) adversary with classical computational resources.
|
|
Can corrupt up to $f < n/3$ validators in a $n$-validator BFT consensus. Can perform
|
|
denial-of-service attacks on network links. Cannot break standard cryptographic
|
|
assumptions (DLP, LWE, collision resistance of SHA-256).
|
|
\end{definition}
|
|
|
|
\textbf{Capabilities}: Software exploits, social engineering, bribery of validators,
|
|
network partitioning, Sybil attacks on peer discovery.
|
|
|
|
\textbf{Limitations}: Cannot compute discrete logarithms. Cannot find SHA-256
|
|
collisions. Cannot break LWE.
|
|
|
|
\subsection{Quantum Adversary}
|
|
|
|
\begin{definition}[Quantum adversary $\mathcal{A}_Q$]
|
|
An adversary with access to a fault-tolerant quantum computer with $\geq 4{,}000$
|
|
logical qubits. Can execute Shor's algorithm (breaking DLP on all deployed elliptic
|
|
curves) and Grover's algorithm (quadratic speedup on unstructured search).
|
|
\end{definition}
|
|
|
|
\textbf{Capabilities}: Everything $\mathcal{A}_C$ can do, plus: recover ECDSA/BLS/EdDSA
|
|
private keys from public keys, decrypt ECDH key exchanges, reduce hash function
|
|
security by $2\times$.
|
|
|
|
\textbf{Limitations}: Cannot break lattice-based assumptions (Module-LWE, Ring-LWE)---
|
|
no known super-polynomial quantum speedup. Cannot break hash preimage resistance
|
|
below $2^{128}$ for 256-bit hashes.
|
|
|
|
\subsection{Economic Adversary}
|
|
|
|
\begin{definition}[Economic adversary $\mathcal{A}_E$]
|
|
A rational adversary who will mount attacks if and only if the expected profit exceeds
|
|
the expected cost. Not bounded by computational limits but by economic incentives.
|
|
\end{definition}
|
|
|
|
\textbf{Capabilities}: Validator bribery (cost: slashable stake), MEV extraction
|
|
(cost: gas fees), governance attacks (cost: token acquisition), oracle manipulation
|
|
(cost: collateral requirements).
|
|
|
|
\textbf{Defenses}: Slashing penalties exceeding potential profit, MEV-resistant
|
|
block construction, governance timelocks, oracle diversity requirements.
|
|
|
|
\subsection{Network-Level Adversary}
|
|
|
|
\begin{definition}[Network adversary $\mathcal{A}_N$]
|
|
An adversary who controls network infrastructure: can delay, reorder, drop, or
|
|
inject messages between nodes. Can partition the network into isolated components.
|
|
Cannot break message authentication (authenticated channels assumed).
|
|
\end{definition}
|
|
|
|
\textbf{Capabilities}: Eclipse attacks (isolating a node), network partitions (splitting
|
|
the validator set), message delay (slowing consensus), traffic analysis (observing
|
|
message patterns).
|
|
|
|
\textbf{Limitations}: Cannot forge authenticated messages. Cannot prevent eventual
|
|
message delivery (partial synchrony assumption).
|
|
|
|
%% ========================================================================
|
|
%% 3. CRYPTOGRAPHIC ASSUMPTIONS
|
|
%% ========================================================================
|