Towards Sound Detection of Virtual Machines

Jason Franklin, Mark Luk, Jonathan M. McCune, Arvind Seshadri, Adrian Perrig, Leendert van Doorn

Abstract: We design, implement, and evaluate a practical timing-based approach to detect virtual machine monitors (VMMs) without relying on VMM implementation details. The algorithms developed in this paper are based on fundamental properties of virtual machine monitors rather than easily modified software artifacts. We evaluate our approach against two common VMM implementations on machines with and without hardware support for virtualization in a number of remote and local experiments. We successfully distinguish between virtual and real machines in all cases even with incomplete information regarding the VMM implementation and hardware configuration of the targeted machine.