Robust Distributed Averaging in Networks

TL;DR

This paper analyzes how adversaries can disrupt distributed consensus in networks through link failures and node value corruption, providing optimal attack strategies under different constraints.

Contribution

It characterizes the optimal adversarial strategies for both link-breaking and node-value corruption attacks in distributed networks.

Findings

Optimal attack strategies are fully characterized for both attack types.

The analysis provides insights into network robustness against adversarial disruptions.

Strategies depend on the attacker's constraints and capabilities.

Abstract

In this work, we consider two types of adversarial attacks on a network of nodes seeking to reach consensus. The first type involves an adversary that is capable of breaking a specific number of links at each time instant. In the second attack, the adversary is capable of corrupting the values of the nodes by adding a noise signal. In this latter case, we assume that the adversary is constrained by a power budget. We consider the optimization problem of the adversary and fully characterize its optimum strategy for each scenario.