DeepDDoS: A Deep-Learning Model for Detecting Software Defined Healthcare IoT Networks Attacks

Abstract

Internet of Things (IoT) brings major security challenges that have prominent social impact. Sensors diversity as well huge amount of generated data represent a big concern for handling security issues. Therefore, companies and organizations are exposed to increasingly aggressive attacks such as ransomware, denial of service (DoS), and distributed denial of service (DDoS). Although IoT devices bring a substantial socio-economic benefits, attacks can create drastically social problems within organizations like hospitals. According to healthcare-based IoT environment, attacks can impact real-time patient data monitoring/collection and consequently effect decision making with respect to critical healthcare IoT devices such as blood pressure, blood sugar levels, oxygen, weight, and even ECGs, etc. In this paper, we propose DeepDDoS, a stable framework that considers deep learning techniques to detect and mitigate, in real time, DoS/DDoS attacks within healthcare-based IoT environment. By leveraging the public available CICDDoS2019 dataset, we show that DeepDDoS outperforms previous studies and achieves a prediction model equals to $98.8\%$. In addition, DeepDDoS architecture gives an enhanced processing delay.