Toward Classification of Phase Change Memory and 3D NAND Flash SSDs Using Power-based Side-channel Analysis in the Time-domain

Toward Classification of Phase Change Memory and 3D NAND Flash SSDs Using Power-based Side-channel Analysis in the Time-domain

연구 분야: Analysis

학회: MEMSYS '22: Proceedings of the 2022 International Symposium on Memory Systems

Intel first introduced their new phase change memory technology (Optane™) to the market in the form of a solid state drive (SSD), which has not been widely investigated for potential vulnerability to side-channel analysis and attack. Like their NAND flash SSD counterparts, the firmware on an Optane SSD has the potential to make it difficult for the user to validate whether operations (e.g., read and write) are being performed as expected on the drive itself. Machine learning-based classification has proven to be a useful tool in validating embedded firmware operations and uncovering unanticipated behavior. In this work, we describe our initial results in an ongoing investigation into the feasibility of using current-draw to analyze the power side-channel in the first commercially available phase change memory technology, the Intel® Optane™. We used time-domain features from our power-based side-channel analysis to compare seven different solid state drives across different technologies (3D XPoint™/Phase Change Memory and 3D NAND Flash), interfaces (Serial "Advanced Technology" Attachment III and Peripheral Component Interconnect Express 3.0 w/ Non-Volatile Memory Express), and form factors (2.5" and M.2), and presented time domain power signatures for both read and write operations. We examined the average power associated with the longest activity pulse for each of these drives and operations, compared them, and highlighted differences across drives of various technologies and interfaces. In doing so, we demonstrated that classifying these drives based on the combination of read and write longest pulse average power is feasible.