Paper2018
Material reviewed: AbstractThe authors use game theory, including static Bayesian games, signaling games, and multi-stage models, to model defender-attacker interactions in cybersecurity under uncertainty and asymmetric information. Per the abstract, the study examines a case involving Advanced Persistent Threats (APTs) and the Tennessee Eastman process, considering that defenders may employ deception techniques such as honey files and honeypots to create strategic uncertainty for attackers.
NetworkOT/ICSHoneypotHoneytoken
Paper2018
Material reviewed: AbstractThe authors propose a game-theoretic model extending signaling games to include a probabilistic deception detector, applied to the design of defensive deception mechanisms in network security. The work analyzes pooling and partially-separating equilibria depending on detector quality.
NetworkAdversary engagement
Paper2018
Material reviewed: AbstractThe authors propose a partially observable stochastic game framework to model interactions between a fully-observing defender and a partially-observing infiltrator in cybersecurity. They formulate the deception strategy problem as a robust mixed-integer linear program and present a scalable approach using parametric Markov decision processes to compute finite-memory infiltration strategies, illustrating the approach on a network security model.
NetworkAdversary engagementMoving target defense
Paper2017
Material reviewed: AbstractThe authors address the difficulty of distinguishing large-scale non-directional attacks from advanced persistent threats (APTs) and propose implementing honeytokens and honeypots as an intrusion detection system by deliberately placing traps within existing environments. They describe deploying a series of interconnected honeytokens, each with an exploitation difficulty level, to categorize attacks, profile the attacker and steer the execution flow into a maze-like scenario aimed at distracting, gathering forensic evidence and potentially identifying or deterring the attacker.
CloudHoneypotHoneytoken
Paper2017
Material reviewed: AbstractThis paper studies a scenario where a defender uses honeynets for active defense to observe an attacker who has penetrated the network. The authors propose an infinite-horizon Markov decision process over a continuous state space to model when to eject the attacker after gathering information, and analyze the robustness of this policy via a Stackelberg game, with simulations on a conceptual network.
NetworkHoneypotAdversary engagement
Paper2015
Material reviewed: AbstractThe authors propose a game-theoretic model for mechanism design in deception-based defense, applied specifically to honeypots in network defense. They analyze cheap-talk games and an extended version with evidence, where the receiver can detect deception with some probability. According to the authors, the model generalizes traditional signaling games and complete-information games, showing numerically that deception detection can eliminate pure-strategy equilibria.
NetworkHoneypotAdversary engagement
Paper—
Material reviewed: AbstractThe thesis proposes an authentication framework combining multi-factor authentication with honeytoken-based deception techniques. It describes three successive architectures: static honeytokens to improve detection of compromised credentials, the 2FHA mechanism combining one-time passwords (Google Authenticator) with honeytokens, and a final architecture with session-bound honeytokens, blockchain-supported auditing, and smart contracts on Ethereum.
IdentityApplicationHoneytokenDecoy