Eavesdropping on Goal-Oriented Communication: Timing Attacks and Countermeasures

TL;DR

This paper addresses timing-based side-channel leakage in pull-based goal-oriented communication for remote Markov process tracking. It develops a formal framework where Eve leverages transmission timings via a forward-backward posterior estimation, and models the interaction as a zero-sum one-sided partially observable game, showing that computing equilibrium is computationally hard. To mitigate leakage, it proposes a hysteresis-based defense that intermittently switches between goal-oriented and periodic scheduling, effectively bounding information leakage while retaining performance gains. Simulations on a 30-state Markov chain demonstrate the trade-offs: periodic scheduling is private but costly in reward, the pure goal-oriented strategy leaks heavily, and the proposed heuristic achieves a controllable balance, with potential extensions to reinforcement learning and push-based updates for broader applicability.

Abstract

Goal-oriented communication is a new paradigm that considers the meaning of transmitted information to optimize communication. One possible application is the remote monitoring of a process under communication costs: scheduling updates based on goal-oriented considerations can significantly reduce transmission frequency while maintaining high-quality tracking performance. However, goal-oriented scheduling also opens a timing-based side-channel that an eavesdropper may exploit to obtain information about the state of the remote process, even if the content of updates is perfectly secure. In this work, we study an eavesdropping attack against pull-based goal-oriented scheduling for the tracking of remote Markov processes. We provide a theoretical framework for defining the effectiveness of the attack and of possible countermeasures, as well as a practical heuristic that can provide a balance between the performance gains offered by goal-oriented communication and the information leakage.

Figures (7)

• Figure 1: The goal-oriented eavesdropping attack: Eve cannot decipher Alice's responses, but gets the timing signal $\tau$.
• Figure 2: Characterization of the optimal scheduling policy as a function of the density decay $\theta$ and the transmission cost $\beta$.
• Figure 3: Information leakage during a single episode, with $\beta=1$, $\theta=32$, and $D=5$. The ADE thresholds $L_{\min}$ and $L_{\max}$ are marked as dashed lines.
• Figure 4: performance as a function of the density decay $\theta$ and the communication cost $\beta$, with $D=5$.
• Figure 5: performance as a function of the density decay $\theta$ and the communication cost $\beta$, with $D=5$.

Theorems & Definitions (2)

• Theorem 1
• Theorem 2