Energy-efficient strategies for mitigating DoS attacks in healthcare IoT systems

Abstract

Edge-cloud continuum integrates edge and cloud computing to facilitate secure and efficient data processing for wide range of applications. It consists of local edge, edge, and cloud computing tiers. To utilize edge-cloud continuum in healthcare, reliable connectivity, low latency, and robust security are significantly important to ensure real-time data communication, security of patient information, and seamless medical operations. Due to the shortcomings of these features in public networks, private 5G network utilization is identified as a promising solution to facilitate critical applications in healthcare. This further increases the demand for edge-cloud continuum as a data-processing architecture in healthcare. In the edge-cloud continuum, the local edge tier brings computing resources near the data source. But the distributive nature and exposure to less-trusted environments create security risks for the local-edge tier. Deploying security solutions such as firewall, encryption, Intrusion Detection Systems (IDS), and threat detection with Artificial Intelligence (AI) requires significant resources (e.g., energy, memory, and storage). Because of the lack of resources in local edge devices, performance degradation occurs when the security algorithms handle threats. Attackers can exploit this vulnerability, especially to deplete the limited energy resources present in the local edge tier, leading to system-level failures in critical Internet of Things (IoT) systems. Even though the existing research extensively examines threat detection and mitigation mechanisms, energy consumption caused by security attacks and defense mechanisms remains an unexplored area of research. Addressing that research gap, this research focuses on providing an energy consumption analysis of Denial of Service (DoS) attacks and a mitigation mechanism for DoS attacks. Firstly, additional energy consumption caused by DoS attack was measured to identify the effect on the energy consumption in the local edge tier. Then, the energy consumption caused by the network layer and application layer rate limiting defense mechanism was evaluated when mitigating the DoS attack. To improve the energy efficiency of application layer defense mechanisms against DoS attack, timer-based and thread-based approaches were experimented and it has shown effectiveness in reducing the energy consumption significantly. Moreover, energy consumption was evaluated for dual- layer network layer and application rate-limiting defense mechanisms against DoS attack. This dual-layer security approach achieved 21.34% average energy efficiency improvement compared to the baseline application layer rate-limiting deployment. This approach is tailored for critical applications with its additional layer of security. This thesis contributes to the understanding of the energy impact of DoS attacks and mitigation mechanisms of DoS attacks. Further, this research provides more energy-efficient security approaches for deployment in healthcare IoT systems.