On the Impossibility of a Perfect Hypervisor (2506.09825v1)

Abstract: We establish a fundamental impossibility result for a `perfect hypervisor', one that (1) preserves every observable behavior of any program exactly as on bare metal and (2) adds zero timing or resource overhead. Within this model we prove two theorems. (1) Indetectability Theorem. If such a hypervisor existed, no guest-level program, measurement, or timing test could distinguish it from native execution; all traces, outputs, and timings would be identical. (2) Impossibility Theorem. Despite that theoretical indetectability, a perfect hypervisor cannot exist on any machine with finite computational resources. These results are architecture-agnostic and extend beyond hypervisors to any virtualization layer emulators, sandboxes, containers, or runtime-instrumentation frameworks. Together they provide a formal foundation for future work on the principles and limits of virtualization.