Boost‐Defence for resilient IoT networks: A head‐to‐toe approach

• Published in: Expert systems Vol. 39; no. 10
• Main Authors: Abu Al‐Haija, Qasem, Al Badawi, Ahmad, Bojja, Giridhar Reddy
• Format: Journal Article
• Language: English
• Published: Oxford Blackwell Publishing Ltd 01.12.2022
• Subjects: Classification; classification methods; cyber security; Datasets; Decision trees; Internet of Things; intrusion detection systems; Machine learning; Networks; New technology; supervised learning
• ISSN: 0266-4720; 1468-0394
• DOI: 10.1111/exsy.12934

The Internet of Things (IoT) is an emerging technology that is considered a key enabler for next‐generation smart cities, industries, security services and economies. IoT networks allow connected devices to communicate with each other automatically without human intervention which empowers innovative solutions for pressing challenges and limitations of current technologies required to materialize smart environments. Due to the concrete involvement of IoT networks in critical infrastructures and cyber‐physical systems, defending them against cyber‐attacks has led to extensive research efforts to propose effective countermeasures against such attacks. In this work, we present Boost‐Defence: a framework to secure IoT networks from a large vector of cyber‐attacks at different IoT layers. We employ the AdaBoost machine learning technique combined with Decision Trees and extensive data engineering techniques to construct a robust classifier for detecting and classifying several cyber‐attacks in IoT networks. We evaluate our system on the TON_IoT_2020 datasets, a collection of datasets compiled specifically for 3‐layered IoT systems comprising: physical, network and application layers. We contrast the performance of our system against existing state‐of‐the‐art solutions. Our experimental analysis demonstrates the capability of our framework in providing superior classification accuracy and lower types 1 and 2 errors for constructing more resilient IoT infrastructures.