Distributed Machine Learning Based Intrusion Detection in IoT

IoT systems face significant security challenges due to their extensive attack surfaces and limited resources. To effectively counter these threats in real- time, intrusion detection algorithms must be both energy-efficient and compatible with the constrained computational capabilities of IoT device...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2024 International Conference on Power, Energy, Control and Transmission Systems (ICPECTS) s. 1 - 4
Hlavní autoři: M, Meena, D, Sathish Kumar, M, Boovin, Doss S, Samuel Karunakara, K, Ganesh, A V, Sanjai
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 08.10.2024
Témata:
Computational modeling
Internet of Things
Intrusion detection
IoT
Logic gates
Machine learning algorithms
MQTT
Protocols
Real-time systems
REST
Security
Supervised learning
Zigbee
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:IoT systems face significant security challenges due to their extensive attack surfaces and limited resources. To effectively counter these threats in real- time, intrusion detection algorithms must be both energy-efficient and compatible with the constrained computational capabilities of IoT devices (including memory, CPU, and storage). While deep learning approaches have proven effective for intrusion detection, their high computational demands make them impractical for resource-limited IoT devices. In this study, we propose developing a hierarchical model for IoT intrusion detection. This model will deploy machine learning (ML) algorithms tailored to the capabilities of IoT devices, while offloading more complex processing tasks to more powerful edge or cloud-based systems. These advanced systems will handle tasks such as updating ML model parameters and executing detection algorithms that exceed the capabilities of the IoT devices. Additionally, our approach will address not only traditional network attacks but also IoT-specific threats related to protocols such as REST, Zigbee, and MQTT.
DOI:10.1109/ICPECTS62210.2024.10780018