IoT Malware Dynamic Analysis Scheme Using the CNN Model

Abstract

Recently, Internet of Things (IoT) technologies have been fused with next-generation technologies such as 5G and deep learning and used in diverse fields such as smart homes, smart cars, and smart appliances. As the demand for IoT devices increases, security threats targeting IoT devices, IoT infrastructure, and IoT application programs have also been increasing. Diverse studies on IoT malware detection have been conducted to protect IoT devices particularly from IoT malware among the security threats. However, existing studies can only accurately detect known IoT malware, not new and variant IoT malware. In this study, the malware dynamic analysis (MALDA) scheme that accurately detects new and variant malware that threatens IoT devices quickly is proposed to reduce the damage caused to IoT devices. The MALDA scheme dynamically analyzes IoT malware in nested cloud environments by training the behavioral features of IoT malware based on the Convolutional Neural Network (CNN) model.