Performance Guarantees for Data Freshness in Resource-Constrained Adversarial IoT Systems

Timely updates are critical for real-time monitoring and control applications powered by the Internet of Things (IoT). As these systems scale, they become increasingly vulnerable to adversarial attacks, where malicious agents interfere with legitimate transmissions to reduce data rates, thereby inflating the age of information (AoI). Existing adversarial AoI models often assume stationary channels and overlook queueing dynamics arising from compromised sensing sources operating under resource constraints. Motivated by the G-queue framework, this paper investigates a two-source M/G/1/1 system in which one source is adversarial and disrupts the update process by injecting negative arrivals according to a Poisson process and inducing i.i.d. service slowdowns, bounded in attack rate and duration. Using moment generating functions, we then derive closed-form expressions for average and peak AoI for an arbitrary number of sources. Moreover, we introduce a worst-case constrained attack model and employ stochastic dominance arguments to establish analytical AoI bounds. Numerical results validate the analysis and highlight the impact of resource-limited adversarial interference under general service time distributions.

💡 Research Summary

The paper addresses the increasingly critical problem of maintaining fresh information in large‑scale Internet‑of‑Things (IoT) monitoring systems that are exposed to resource‑constrained adversarial interference. The authors model a two‑source M/G/1/1 queue in which a legitimate sensor (source s₁) generates status updates according to a Poisson process with rate λ₁, while a compromised sensor (source s₂) injects “negative” arrivals also according to a Poisson process with rate λ₂. A negative arrival does not carry useful data; instead, it preempts any update currently in service, forcing the server into a slowdown mode where the service time of the preempted packet is multiplied by a factor β > 1. The slowdown persists for an exponentially distributed period with mean 1/Λ, where Λ = λ₁ + λ₂, after which normal service resumes. To keep the attack model realistic, the adversary’s rate is bounded (λ₂ ≤ λ_max) and the slowdown factor is limited (β ≤ β_max), reflecting battery, spectrum, or detection constraints.

The analysis proceeds by first characterizing the system time T₁ of a successfully delivered update from the legitimate source. Conditioning on the event D that the update finishes service before any negative arrival yields a probability Pr(D) = M_{Sₙ}(−λ₂), where M_{Sₙ}(·) is the moment‑generating function (MGF) of the nominal service time Sₙ. This leads to a closed‑form pdf f_{T₁}(t) = f_{Sₙ}(t) e^{−λ₂t}/M_{Sₙ}(−λ₂) and an MGF M_{T₁}(s) = M_{Sₙ}(s − λ₂)/M_{Sₙ}(−λ₂).

Next, the inter‑departure time Y₁ (the time between two consecutive successful deliveries) is modeled as a semi‑Markov process with three states: q₀ (idle, waiting for a new positive packet), q₁ (normal service), and q₂ (slowdown service). The authors define sojourn times W₁…W₅ for each transition, derive their MGFs (e.g., E