cloud-hypervisor/cloud-hypervisor
Cloud Hypervisor
Cloud Hypervisor is a Rust-based hypervisor and KVM virtual machine monitor designed to execute 64-bit guest operating systems. It functions as a user-space virtual machine manager that employs a minimal emulation layer to reduce memory overhead and latency for cloud workloads.
The project distinguishes itself through the use of a memory-safe language to implement a virtio device emulator and a user-space device model. It provides a standardized web API for managing virtual machine lifecycles and resource configurations.
The platform covers broad virtualization capabilities, including the emulation of NVMe and block storage, network connectivity via host bridging, and hardware device passthrough. It supports high-availability operations such as live migration, state snapshotting, and the dynamic resizing of CPU and memory resources through hotplugging.
The system is managed via a REST-API control plane and provides secure communication channels and shared memory interfaces between the host and guest.
Features
- Cloud Workload Virtualizers - Functions as a minimal emulation layer designed to execute 64-bit guest operating systems with optimized resource usage for cloud workloads.
- Cloud Workload Execution - Executes 64-bit guest operating systems using a minimal emulation layer to reduce memory usage and latency.
- QEMU KVM - Uses the Linux kernel virtual machine (KVM) interface to execute guest operating systems with near-native performance.
- Block Storage Services - Exposes virtual block devices to allow a guest operating system to boot and store data.
- Control Plane APIs - Manages virtual machine lifecycles and resource configurations through a standardized REST-API control plane.
- VM Live Migration - Moves running virtual machines between physical hosts with minimal interruption to ensure continuous service and resource balance.
- Virtual Machine Lifecycle Management - Manages the lifecycle of virtual machines through web endpoints, including dynamic CPU and memory resizing and power actions.
- Cloud Workload Virtualization - Runs 64-bit guest operating systems with a minimal emulation layer to reduce memory overhead and latency for cloud environments.
- User-Space Device Models - Runs hardware emulation and device state management in user space to improve security and stability.
- Rust-Based Hypervisors - Utilizes the Rust programming language to ensure memory safety and high performance in the hypervisor's implementation.
- User-Space Virtual Machine Managers - Manages guest CPU and memory resources as a user-space process via a standardized web API.
- Virtio Device Emulators - Provides an implementation of paravirtualized I/O devices for networking and storage using the virtio standard.
- Hardware Passthrough - Maps physical host PCI devices directly into the guest address space using an IOMMU for high-throughput hardware access.
- Virtual Machine Migration - Moves running virtual machines between physical hosts to ensure high availability and resource balance.
- Virtual Machines - Implements a lightweight virtual machine monitor built on KVM for running isolated guest environments.
- Live Migrations - Serializes the guest memory and CPU state to transfer a running virtual machine between physical hosts.
- Virtio Device Emulation - Implements standardized paravirtualized interfaces to handle network, block, and console I/O between the host and guest.
- Virtual Machine Snapshots - Saves the current execution state of a virtual machine to disk to resume operation from that point.
- Networking and Connectivity - Manages virtual network isolation and connectivity for guests using virtio-net or vhost-user-net backends.
- Host-Guest Mounts - Maps host directories into virtual machine filesystems for seamless file access via virtio-fs.
- Shared Memory IPC - Uses memory-mapped regions and virtio queues to exchange data between the host and guest without expensive context switches.
- Host-Guest Communication Channels - Implements a secure communication channel between host and guest using a hybrid socket address family over virtio.
- Dynamic CPU Resizing - Allows adjusting the number of virtual CPUs in a running instance dynamically via API requests.
- Dynamic Memory Resizing - Expands or shrinks available RAM in a running instance using memory management protocols via API requests.
- Hardware Resource Hotplugging - Enables adding or removing CPUs, memory, and PCI devices while the guest system remains running.
- Hardware Resource Hotplugging - Updates the guest hardware configuration by injecting virtual PCI devices or memory blocks into a running instance.
- DMA and IOMMU Translations - Implements a virtual input output memory management unit to enable safe nested virtualization and device assignment.
- Persistent Memory Emulation - Emulates persistent memory devices to reduce memory footprint and bypass guest page caches.
- Physical Device Pass-through - Implements the mapping of physical host PCI devices directly into the guest address space for high-throughput hardware access.
- Virtual Disk Image Booting - Starts virtual machines directly from cloud images or formatted kernels to streamline deployment.
- NVMe Emulation - Simulates NVMe controllers to provide high-speed storage access to the virtual environment.
- Virtual Storage Emulation - Provides high-speed access to guests through simulated NVMe controllers, block devices, and shared host filesystems.
- Host-Guest Integration Tools - Establishes secure communication channels and shared memory interfaces between a host and guest using virtio protocols.
- Virtual Device Hotplugging - Adds or removes hardware components from a running virtual machine without requiring a system reboot.
- PCI Device Hotplugging - adds or removes PCI hardware from a running instance using unique device identifiers and API requests.
- Network Bridges - Connects virtual machines directly to the host network using a MACvTAP bridge for external connectivity.
- Management APIs - Provides a set of standardized web endpoints to manage and monitor running virtual machines.
- Virtualization and Orchestration - Virtual machine monitor optimized for cloud workloads.
- Hypervisor Implementations - VMM optimized for cloud-native workloads.
