Securing cloud storage systems through a virtual machine monitor
Fatemeh Azmandian, David R Kaeli, Jennifer Dy, Javed A. Aslam, Dana Schaa · 2012
Cloud storage solutions have increasingly gained in popularity as they offer a convenient method of maintaining one's data all in one place, with the ability to access it from anywhere at any time. In this paper, we leverage a virtualization-based intrusion detection infrastructure to build secure cloud storage systems. The intrusion detection system uses machine learning techniques applied to data available at the virtual machine monitor layer to identify the presence of malicious activity during a workload's execution. Our results show that by running a cloud storage server in a virtual execution setting, we can detect real-world malware attacks at a high detection rate of over 98%, with fewer than 3% false alarms.