Comparative Analysis of Malware Detection Approaches in Cloud Computing

The widespread use of cloud computing techniques in many applications renders cloud computing environments extremely vulnerable to malware infections and novel attacks.The flexibility, scalability, and elasticity cloud computing provide add to the difficulty of detecting malicious software in cloud computing environments.In this study, we analyze malware attacks that infect cloud computing environments.Moreover, we elaborate on different malicious software detection approaches in cloud computing environments.Furthermore, we evaluate these approaches by considering other perspectives (i.e., malware detection accuracy and deployed analytical techniques).More than 50% of the approaches of the malware detection papers (in this survey) used deep learning techniques in cloud computing environments.In addition, the majority of authors preferred to use dynamic malware analysis.Deep learning and dynamic analysis are powerful, complementary approaches in malware detection.Dynamic analysis observes the runtime behavior of programs, such as API calls, file operations, and network activity, to detect malicious patterns in controlled environments like sandboxes.When integrated with deep learning, this behavioral data can be analyzed more effectively using advanced models like RNNs or CNNs.Deep learning enhances dynamic analysis by identifying complex, hidden patterns in malware behavior and adapting to zero-day threats.This combination provides a robust defense mechanism, particularly in cloud computing, where large-scale and real-time detection capabilities are critical.The rates of detection are vacillated from 79% to 99%.