Carbon-Aware Intrusion Detection: A Comparative Study of Supervised and Unsupervised DRL for Sustainable IoT Edge Gateways

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Jamshidi, Saeid, Khomh, Foutse, Nafi, Kawser Wazed, Nikanjam, Amin, Keivanpour, Samira, Abdul-Wahab, Omar, Bellaiche, Martine
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914581046099968
author Jamshidi, Saeid
Khomh, Foutse
Nafi, Kawser Wazed
Nikanjam, Amin
Keivanpour, Samira
Abdul-Wahab, Omar
Bellaiche, Martine
author_facet Jamshidi, Saeid
Khomh, Foutse
Nafi, Kawser Wazed
Nikanjam, Amin
Keivanpour, Samira
Abdul-Wahab, Omar
Bellaiche, Martine
contents The rapid expansion of the Internet of Things (IoT) has intensified cybersecurity challenges, particularly in mitigating Distributed Denial-of-Service (DDoS) attacks at the network edge. Traditional Intrusion Detection Systems (IDSs) face significant limitations, including poor adaptability to evolving and zero-day attacks, reliance on static signatures and labeled datasets, and inefficiency on resource-constrained edge gateways. Moreover, most existing DRL-based IDS studies overlook sustainability factors such as energy efficiency and carbon impact. To address these challenges, this paper proposes two novel Deep Reinforcement Learning (DRL)-based IDS: DeepEdgeIDS, a label-free Autoencoder-DRL hybrid, and AutoDRL-IDS, a supervised LSTM--DRL model. Both DRL-based IDS are validated through theoretical analysis and experimental evaluation on edge gateways. Results demonstrate that AutoDRL-IDS achieves 94% detection accuracy using labeled data, while DeepEdgeIDS attains 98% offline evaluation accuracy through label-free anomaly detection and online mitigation feedback. This study introduces a carbon-aware, multi-objective reward formulation that supports supervised reward optimization for AutoDRL-IDS and label-free online reward learning for DeepEdgeIDS, enabling sustainable real-time IDS operation in dynamic IoT networks.
format Preprint
id arxiv_https___arxiv_org_abs_2511_18240
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Carbon-Aware Intrusion Detection: A Comparative Study of Supervised and Unsupervised DRL for Sustainable IoT Edge Gateways
Jamshidi, Saeid
Khomh, Foutse
Nafi, Kawser Wazed
Nikanjam, Amin
Keivanpour, Samira
Abdul-Wahab, Omar
Bellaiche, Martine
Cryptography and Security
The rapid expansion of the Internet of Things (IoT) has intensified cybersecurity challenges, particularly in mitigating Distributed Denial-of-Service (DDoS) attacks at the network edge. Traditional Intrusion Detection Systems (IDSs) face significant limitations, including poor adaptability to evolving and zero-day attacks, reliance on static signatures and labeled datasets, and inefficiency on resource-constrained edge gateways. Moreover, most existing DRL-based IDS studies overlook sustainability factors such as energy efficiency and carbon impact. To address these challenges, this paper proposes two novel Deep Reinforcement Learning (DRL)-based IDS: DeepEdgeIDS, a label-free Autoencoder-DRL hybrid, and AutoDRL-IDS, a supervised LSTM--DRL model. Both DRL-based IDS are validated through theoretical analysis and experimental evaluation on edge gateways. Results demonstrate that AutoDRL-IDS achieves 94% detection accuracy using labeled data, while DeepEdgeIDS attains 98% offline evaluation accuracy through label-free anomaly detection and online mitigation feedback. This study introduces a carbon-aware, multi-objective reward formulation that supports supervised reward optimization for AutoDRL-IDS and label-free online reward learning for DeepEdgeIDS, enabling sustainable real-time IDS operation in dynamic IoT networks.
title Carbon-Aware Intrusion Detection: A Comparative Study of Supervised and Unsupervised DRL for Sustainable IoT Edge Gateways
topic Cryptography and Security
url https://arxiv.org/abs/2511.18240