Using Process Mining to Identify File System Metrics Impacted by Ransomware Execution

Malware authors leverage strong cryptographic primitives to hold user files as a hostage in their own devices until a ransom is paid. Indeed, victims not protected against ransomware are forced to pay the ransom or lose the files if ignoring the extortion. Devices are by no means immune from ransomware attacks. The reality is that there is a limited study on how to protect end-user devices against ransomware while there is hardly any protection available. Ransomware uses legiti- mate operating system processes that even state-of-the-art and advanced anti-malware products are ineffective against them. The results of our static and dynamic analysis illustrate that a local file system plays a critical role in the operation of all ransomware engines. Therefore, this study investigates the correlation existed between the file system opera- tions to identify metrics such as the absolute occurrence frequency of a system file to identify a ransomware attack from within the kernel. We employ business process mining techniques to analyze collected log files from samples of seven recent live ransomware families and use the Naive discovery algorithm to study the absolute occurrence frequency of system files. The findings are visualized by state charts and sequence diagrams. Finally, the study identifies eight common system files that ransomware calls on in order to encrypt a victim’s files on their device.