Linux & Open Source · Infrastruttura · YoctoIT tech page

KVM

The hypervisor inside the kernel: the open base of OpenShift Virtualization, Proxmox and the clouds — the concrete alternative.

FOCUS · THE OPEN HYPERVISORKVM is everywhere: in the clouds, in OpenShift, in Proxmox — and maybe in your future
YoctoIT material for clients and partners · Linux is a trademark of Linus Torvalds; the distributions and products mentioned are trademarks of their respective owners.
01 · What it is

Virtualizzazione KVM, made clear.

KVM is the hypervisor inside the Linux kernel: it runs the big clouds and sits under the concrete alternatives to VMware — OpenShift Virtualization for those going toward containers, Proxmox VE for those wanting simple, free virtualization. Near-native performance, a mature ecosystem (libvirt, QEMU) and zero hypervisor fees.

In-kernel
the hypervisor is the kernel: performance and security from Linux's heart
Ovunque
AWS, Google, OpenStack: the clouds have always run on KVM
Proxmox
the KVM+containers distribution conquering the SMBs
Virtualizzazione KVM
OFFICIAL LINUX BRANDING · KVM
CONSOLE REALE · PROXMOX VE · FONTE: PROXMOX
REAL CONSOLE · PROXMOX VE · SOURCE: PROXMOX
02 · How to use it well

The things that make the difference.

The incarnations

The usual VMsWindows and Linux, everywhere
Proxmox VE
OpenShift Virt
pure libvirt
simple · integrated · custom
Ceph/ZFS & SDNopen storage and network alongside
KVM in the kernelthe common engine
One engine, three bodies

The post-Broadcom evaluation

Proxmox or OpenShift Virt instead of vSphere? The PoC with YOUR workloads: the answers from facts.

VM migration

virt-v2v and conversions from the VMware formats: the move rehearsed, with rollback.

HA and backups done well

Proxmox clusters, fencing and the pairing with Veeam/PBS: serious production on the open too.

The skills underneath

KVM rewards those who know the Linux underneath: that's our case — the advantage is yours.

03 · In depth

KVM: the hypervisor that's already in the kernel

KVM is the hypervisor inside the Linux kernel: QEMU emulates, libvirt governs, virt-manager and Cockpit give the UIs; the performance is bare-metal grade (VirtIO for paravirtualized I/O, hugepages, CPU pinning and NUMA for the serious workloads), the live migration moves the VMs hot, SEV/TDX encrypt the VMs' memory; it's the base of Proxmox, OpenStack, oVirt and Red Hat Virtualization: the concrete post-Broadcom alternative, without per-socket license costs.

  • Nel kernel — no hypervisor to install: Linux IS the hypervisor
  • VirtIO — the paravirtualized I/O: performance close to the metal
  • Live migration — the VMs moved hot: maintenance without downtime
  • NUMA & pinning — the tuning for the databases: the VM that doesn't pay for the virtualization
  • SEV/TDX — the VM memory encrypted: confidential computing
  • Zero licenze — no per-socket toll: the TCO that changes the conversation
04 · Numbers and lifecycle

The numbers that matter.

~98%
the performance vs bare metal with VirtIO
0€
the hypervisor's cost: it's in the kernel
s
the live migration on a fast network
upstream
the development: the worldwide kernel community
KVM is mature for the enterprise: the post-Broadcom evaluation we do with numbers — proof, migration and run.
05 · Use cases

Where it really pays off.

An alternative to VMware

The hypervisor fee zeroed, with judgment.

Labs and DR

The secondary environments, free and functional.

Custom infrastructures

Where total control of the stack is needed.

The open hypervisor has been mature for years: the question is the surroundings — and the surroundings are our trade.