Date of Publication :21st May 2024

Abstract:Malicious software or malware is on the rise, with an increasing number of sophisticated variants employing various obfuscation techniques. Detecting malware before it wreaks havoc on computer systems and the Internet is imperative. This paper offers a comprehensive survey of existing malware detection approaches, shedding light on the persistent challenges in this domain. This paper presents a literature survey that delves into cloud-based malware detection methods or models. This thorough analysis looks at four different methods for detecting malware, each designed for a particular environment and taking care of a different set of problems. The Proposed Malware Detection Model (PMDM) and Cloud Deployment Model (CDM) make up the first model, which suggests a malware detection system for cloud deployment. PMDM uses behavioral, symbolic, and DNA sequence detection processes to improve system flexibility and speed. To detect malware in a scalable and easily accessible manner, CDM uses Eucalyptus in an actual cloud setting. The second method uses a Cloud-Based Behavior Centric Model, dynamic analysis tools, and several machine learning algorithms to present an intelligent behavior-based malware detection system for cloud environments. The third technique, called TrustAV, maximizes malware scanning efficiency by utilizing a multimodal strategy inside Intel SGX enclaves. The fourth framework uses the computational power of security labs to simulate end- user environments and presents a cloud-based malware analysis system based on dynamic behavior. A thorough examination of system call interception and proxying, one-way isolation, and integration with currently available malware detectors are included in the paper’s conclusion. This review sheds light on various approaches, their advantages, disadvantages, and possible areas for development in the dynamic field of intelligent malware detection.

Reference :