Deciphering Ransomware: Strategic API Usage and Behavioral Patterns for Advanced Detection Techniques

Résumé

Ransomware has emerged as a critical cybersecurity threat, inflicting severe financial and operational damage across industries. Traditional signature-based detection systems struggle to detect zero-day and evolving ransomware strains, as they rely on known signatures that cannot capture new tactics. In contrast, behavioral detection methods analyze ransomware actions and patterns, making them more effective. Integrating artificial intelligence (AI) can further improve detection rates; however, effective AI models require diverse, up-to-date, and representative data. Previous research has often focused on isolated aspects of ransomware behavior. Our study addresses these gaps by providing a publicly accessible, up-to-date dataset covering multiple ransomware variants and families, including polymorphic and obfuscated strains not comprehensively explored in prior literature. Additionally, our approach identifies extensive ransomware metrics, including network interactions, registry modifications, file system changes, and low-level API call patterns, enabling real-time detection of malicious activities. Through a comprehensive behavior-based analysis of over 200 recent ransomware samples using the Cuckoo Sandbox platform and custom Python scripts, our study provides cybersecurity practitioners with valuable data and actionable insights, supporting faster responses, improved threat detection, and a proactive stance against evolving risks.

Citer ce document

Accès au document

Texte intégral en lecture en ligne, réservé aux abonnés SPHAERO et aux membres de l'institution. Se connecter

Auteur(s)

Statistiques

Consultations : 1

Téléchargements : 0