Towards a Novel Approach for Hidden Process Detection Based on Physical Memory Scanning

摘要

Leveraging developed root kit, malware could deeply hide its own process and hardly be detected. Based on analyzing various existing detecting technologies, a novel approach for hidden process detection was proposed in this paper. The approach used page table entry patching to traverse physical memory and obtain the raw data, and formulated the characteristic selection constraints to extract reliable process object characteristics, which were used to search process object instances based on string matching in physical memory to form a credible list of processes. The approach could also be used to search other kernel objects on varieties of system platforms. The experimental results show that new detection is effective in hidden process searching.