paravirtualization

Paravirtualization is a virtualization technique in which the guest operating systems are modified to run on a virtual machine monitor (hypervisor) by replacing privileged instructions with hypercalls and by using specialized drivers to access hardware resources. In this approach, the guest cooperates with the hypervisor rather than trapping and emulating every privileged operation.

The guest OS communicates with the hypervisor through hypercalls for actions such as memory management, scheduling, and I/O. The hypervisor presents a virtual hardware interface, and pV drivers (paravirtualized drivers) are used to perform I/O and other operations efficiently. This cooperation reduces the overhead of virtualization compared with full trapping of privileged instructions.

Historically, paravirtualization gained prominence with the Xen hypervisor in the mid-2000s. It requires modifying the guest OS to incorporate PV components. Linux and several BSD variants have supported PV builds, and Windows guests could run in paravirtualized mode when vendors provided PV drivers. With hardware-assisted virtualization (Intel VT-x/AMD-V), the benefit of paravirtualization diminished for some workloads, leading to hybrids such as PVH that blend paravirtualization with hardware-assisted techniques.

Pros of paravirtualization include lower overhead and improved I/O performance on hardware without robust virtualization acceleration, provided the OS is modified accordingly. Cons include reduced portability and increased maintenance burden due to dependence on OS-specific PV components and versions.

Today, paravirtualization is less common as a primary approach, having largely evolved into hybrid or driver-based models within modern hypervisor ecosystems, while preserving its influence on efficient I/O virtualization.