Strengthening Cybersecurity in Edge Computing with Machine Learning

Abstract

Edge computing is an emerging revolution that decentralizes computational tasks from traditional data center platforms to bring analytics algorithms closer to the source of data. However, this change in the paradigm invokes new security concerns that require modern protection methods to protect the Edge environment. In this paper, I study the incorporation of machine learning (ML) solutions to improve edge computing functionality for better cybersecurity. Specifically, we look into different ML algorithms and how they can be used to identify edge security threats like intrusion, data breaches, or malicious actions. Our study shows the efficacy of ML-driven solutions in sensing and mitigating real-time, emerging threats with automated anomaly detection using a behavior analysis platform that can leverage adaptive learning to proactively respond. We present empirical analysis and case studies to discuss the pros and cons of existing ML methods in edge applications. We learned that ML is a game changer for threat detection and response, but you have to deal with the same challenges around it: resources are tight and models do not scale well. We ended our paper with an outlook on securing edge computing using ML strategies, recommendations for further optimization of these strategies at the Machine-to-Machine (M2M) level, and potential future research areas addressing existing issues utilizing this particular domain.