paravirtualizace

Paravirtualization is a technique in virtualization that involves modifying the guest operating system's kernel to make it aware it is running in a virtualized environment. This awareness allows the guest OS to communicate directly with the hypervisor through a paravirtualized interface, often referred to as a "hypercall." Unlike full virtualization, where the hypervisor intercepts and emulates all hardware calls from the guest, paravirtualization relies on the guest OS to make specific, optimized calls to the hypervisor for privileged operations.

This approach offers significant performance advantages because it bypasses the overhead associated with hardware emulation. The guest OS is essentially cooperating with the hypervisor, leading to faster execution of I/O operations, system calls, and other resource-intensive tasks. However, a major drawback of paravirtualization is that it requires modifications to the guest operating system. This means that unmodified, off-the-shelf operating systems cannot be directly run with a paravirtualized hypervisor without some form of adaptation or binary translation.

Examples of hypervisors that utilize paravirtualization include Xen, which historically relied heavily on it, and KVM, which can employ paravirtualized drivers (like virtio) for improved performance alongside full virtualization capabilities. Paravirtualization has been crucial in achieving near-native performance in certain virtualization scenarios, especially for workloads that are sensitive to I/O latency.