cloudparts

ACRNDormant

A small type-1 hypervisor for Intel-based embedded and industrial systems that runs real-time and safety VMs on dedicated cores next to Linux, Windows or Android guests on the same board.

Linux Foundation project
What it is
Hypervisors
Type
Open source
Made by
Project ACRN (Linux Foundation), started by Intel
Foundation
The Linux Foundation, joined March 2018. A series of LF Projects, LLC, started with code from Intel
License
BSD-3-Clause
Status
Dormant. No release since v3.3 in August 2024, and no commits to the main branch since November 2025.
Facts checked
28 October 2026

What it is

ACRN (pronounced “acorn”) runs on bare metal using Intel VT and is about 40,000 lines of code; its documentation compares that with more than 150,000 for data-centre hypervisors. It is built to consolidate workloads that used to need separate devices, such as soft PLCs, robotics control and general-purpose desktops, onto one Intel-based industrial or automotive computer. Pre-launched VMs are started by the hypervisor itself at boot with their own CPU cores, memory and devices, and run independently of everything else: a real-time VM (RTVM) running Zephyr, VxWorks or Xenomai for latency-critical control, or a Safety VM that watches for hardware failures and takes emergency action. A Service VM, normally Ubuntu, owns most of the remaining hardware and runs the Device Model, a user-space program that emulates devices and launches post-launched User VMs (Ubuntu, Windows, Android through Intel’s Celadon) that share that hardware through it.

Three scenarios combine these. A shared scenario has a Service VM and post-launched VMs; a partitioned scenario has only pre-launched VMs, with no Service VM or Device Model; a hybrid scenario mixes the two. The layout is fixed before build time: the Board Inspector records the target hardware in a board XML file, the ACRN Configurator turns that into a scenario XML file of VMs, CPU affinity and devices, and the hypervisor is compiled for that scenario and installed as Debian packages. ACRN was announced as a Linux Foundation project in March 2018 with ADLINK, Aptiv, Intel, LG Electronics and Neusoft as early members, and TTTech and ADLINK shipped products on it. Activity has since faded: v3.3 in August 2024, which requires Ubuntu 22.04 on the target, is the latest release, the main branch had 26 commits in 2025 and none so far in 2026, and the project website’s newest dated news is from 2021.

Why teams pick it

  • Small hypervisor: about 40,000 lines of code, a fraction of a general-purpose data-centre hypervisor, which keeps the trusted code reviewable
  • Mixed criticality on one board: real-time and safety VMs get dedicated cores, memory and devices next to rich guests such as Ubuntu, Windows or Android
  • Partitioned mode needs no Service VM at all, so a failure in a general-purpose OS cannot touch the pre-launched VMs
  • Permissive BSD-3-Clause licence for the hypervisor and Device Model, which suits embedded products that ship modified code

Watch out for

  • Little ongoing development: no release since August 2024 and no main-branch commits in 2026, so newer hardware, kernels and Ubuntu releases may not be supported
  • Intel platforms only: it relies on Intel VT and runs in VMX root mode, and its documentation and tools target Intel boards
  • Static configuration: the board and scenario XML files are compiled into the hypervisor, so changing VMs or CPU assignments means rebuilding and reinstalling it, and the Board Inspector must run on the target with the BIOS already set as it will stay
  • Tooling is tied to one OS version: the current guide needs Ubuntu 22.04 on both the development computer and the target, and development on Windows isn’t supported

Alternatives

What all hypervisors have in common →Compare side by side →
Xen Project hypervisorOpen sourceThe open-source type-1 hypervisor from Cambridge that ran the first Amazon EC2 instances: a small microkernel-style core that boots first, with drivers and management in a privileged VM called dom0.KVM (Kernel-based Virtual Machine)Open sourceThe hypervisor built into the Linux kernel: it runs guest code directly on the CPU through /dev/kvm, and leaves device emulation to a user-space program such as QEMU or Firecracker.Cloud HypervisorOpen sourceA Rust virtual machine monitor for modern cloud VMs on KVM or Microsoft’s hypervisor: virtio devices only, plus hotplug, VFIO passthrough, live migration, snapshots and Windows guests.HeliosOpen sourceOxide’s distribution of illumos, the operating system on every sled’s host CPU in the Oxide rack, where it runs VMs on its bhyve hypervisor and the control plane in zones.libvirtOpen sourceA stable API, daemon and virsh command line for managing virtual machines, storage and networks, with drivers for QEMU/KVM, Xen, LXC, Cloud Hypervisor, VMware ESX, Hyper-V and more.Microsoft Hyper-VCommercialMicrosoft’s type-1 hypervisor, built into Windows Server and Windows 10 and 11 Pro and Enterprise: a thin hypervisor under a Windows root partition, which also underpins Azure’s hypervisor and Windows features like VBS, WSL 2 and Windows Sandbox.Oracle VirtualBoxOpen sourceOracle’s free, open-source desktop hypervisor: a type-2 hypervisor that runs VMs as an application on Windows, macOS, Linux and Solaris, scriptable with VBoxManage and the default provider for Vagrant.PropolisOpen sourceOxide’s virtual machine monitor: a Rust program that pairs with the bhyve hypervisor in the illumos kernel and emulates the devices each VM in an Oxide rack sees.Proxmox Virtual EnvironmentOpen sourceA Debian-based server platform from Vienna that runs KVM virtual machines and LXC containers, with clustering, high availability, Ceph and ZFS storage, backups and a web interface in one free ISO.QEMUOpen sourceThe open-source machine emulator and virtualizer: it emulates whole computers in software, or runs them at native speed on a hypervisor such as KVM, and supplies the virtual devices either way.VMware ESXiCommercialVMware’s bare-metal hypervisor (called ESX again since VCF 9), whose own kernel runs VMs directly on servers while vCenter adds vMotion, HA, DRS and vSAN; since Broadcom bought VMware in 2023 it is sold only as per-core subscriptions.

Scenarios#

Chosen in the ACRN Configurator and built into the hypervisor.

Shared

Service VM + User VMs

The Service VM owns the hardware and its Device Model launches User VMs that share it; the most flexible layout, and the one the getting-started guide builds.

Partitioned

pre-launched VMs only

Each VM is started by the hypervisor with its own CPUs, memory and passed-through devices. There is no Service VM or Device Model, and nothing shared to fail.

Hybrid

both

Pre-launched VMs, such as an RTVM or a Safety VM, run beside a Service VM that launches shared User VMs, so control and HMI workloads sit on one board.

See it in a real stack

No architecture uses ACRN yet. Drop it into the playground and wire it up yourself.Open playground →
Where it came from2018–2025

Embedded systems such as industrial controllers and car cockpits were moving to multicore Intel chips that could replace several separate boxes, if something could keep a real-time control loop or a safety monitor isolated from a general-purpose OS on the same chip. Data-centre hypervisors were large and not designed for real-time or safety work. Intel wrote ACRN to fill that gap, and on March 13, 2018, at the Embedded Linux Conference in Portland, the Linux Foundation announced it as a “flexible, lightweight reference hypervisor, built with real-time and safety-criticality in mind”, with ADLINK, Aptiv, Intel, LG Electronics and Neusoft as early members.

Releases came every month or two at first: v0.1 in July 2018, 1.0 in May 2019 and 2.0 in June 2020, which added the hybrid mode for running safety-critical and resource-sharing VMs together. TTTech Industrial shipped ACRN 2.0 in its Nerve Blue edge platform in December 2020, and ADLINK announced support in its ROScube-I ROS 2 robot controller in January 2021. The 3.x line from June 2022 slowed to one release a year, ending with v3.3 in August 2024; commits to the main branch dropped from 144 in 2024 to 26 in 2025, the last in November 2025.

Why it’s built this way

A small hypervisor with devices in a Service VM
Keeping drivers and device emulation in the Device Model, a program in a Linux Service VM, kept the hypervisor near 40,000 lines, small enough to aim at real-time and safety use.
Pre-launched VMs with dedicated hardware
Real-time and safety workloads can’t wait on a general-purpose OS to boot or share a device with it. Starting them from the hypervisor with their own cores, memory and devices keeps them independent of the Service VM.
Configure at build time, not run time
Fixing the board and the VM layout in XML compiled into the hypervisor makes the system’s behaviour predictable and removes code for changing it while running, at the cost of rebuilding for each change.

Timeline

  1. 2018

    Announced as a Linux Foundation project at ELC (March 13); v0.1 in July

  2. 2019

    ACRN 1.0 released (May)

  3. 2020

    ACRN 2.0 adds hybrid mode (June); TTTech ships it in Nerve Blue (December)

  4. 2022

    ACRN 3.0 released (June)

  5. 2024

    ACRN 3.3, the latest release (August)

  6. 2025

    Last commit to the main branch (November)

Sources8 · facts checked 28 October 2026

Improve this page on GitHub