Cloud data security for distributed embedded systems using machine learning and cryptography

In the growing demand for distributed embedded systems that efficiently execute complex processes and high-end applications, safeguarding sensitive data is imperative. The landscape of security threats, cyber-attacks, and associated challenges has reached unprecedented sophistication and magnitude. Whether you are an individual, an organization, or a company invested in distributed embedded systems, your paramount concern revolves around protecting invaluable data. In this digital age, data security is genuinely where the wealth resides. This paper introduces a novel framework for securing the sensitive data of distributed embedded systems. The framework comprises five stages, each contributing to an efficient and robust data security approach. Additionally, this work proposes developing an efficient Multiple Attack Detection model using a supervised machine-learning system. The targeted cyber-attacks include phishing, malware, Distributed Denial-of-Service (DDoS), and DNS over HTTPS attacks. Each targeted cyber-attack is trained on eight supervised machine learning classifiers. The classifier exhibiting the best overall performance in evaluation metrics (accuracy, recall, precision, and F1-score) is integrated into the proposed Multiple Attack Detection model. Furthermore, this paper explores deploying a machine learning-based Multiple Attack Detection model to a cloud environment using Streamlit Cloud.