Cloud Hypervisor 48.0, a hypervisor based on components of the Rust-VMM project developed with the participation of Intel, Alibaba, Amazon, Google, and Red Hat, has been released. Rust-VMM is written in the Rust language and allows for the creation of task-specific hypervisors. Cloud Hypervisor is one such hypervisor, providing a high-level virtual machine monitor (VMM) running on top of KVM or MSHV and optimized for tasks specific to cloud systems. The project code is available under the Apache 2.0 and BSD licenses. The hypervisor was originally created by Intel and has been brought under the auspices of the organization. Linux Foundation, after which Alibaba, AMD, Ampere, ARM, ByteDance, Cyberus Technology, Microsoft, and Tencent Cloud joined the development.
Cloud Hypervisor is focused on running modern distributions Linux Using paravirtualized devices based on Virtio. Key objectives include high responsiveness, low memory consumption, high performance, simplified configuration, and reduced attack vectors. It features the ability to directly boot the OS kernel without intermediate bootloaders, allowing for extremely fast virtual machine startup (less than 100 ms before control is transferred to user-space applications).
Emulation support is kept to a minimum, with a focus on paravirtualization. The project supports x86_64, AArch64, and RISC-V architectures. Guest systems support 64-bit builds. Linux ΠΈ Windows 10/Windows Server 2019, and also includes the ability to use Cloud Hypervisor to run containers in virtual machines using Kata Containers. Live migration is also supported. virtual machines between servers and virtual machine lifecycle management using HTTP API.
In the new version:
- The rules for the use of AI tools in development prohibit the acceptance of code generated using large language models and AI assistants such as ChatGPT, Gemini, Claude, and GitHub Copilot. The stated reason is to avoid ambiguity with compliance with licensing requirements (AI is trained on code under different licenses and the generated code can potentially be treated as a derivative work). The ban on AI is also due to the desire to reduce the burden on maintainers by relieving them of the need to analyze meaningless patches created using AI.
- Added experimental support for the fw_cfg (Firmware Configuration) device, which allows data and settings such as virtual machine boot configuration, system memory layout information, and ACPI tables to be passed from the host environment to the guest system.
- Added experimental support for the ivshmem device to enable shared access to a common memory area between multiple guests.
- Added support for booting RISC-V systems via Firmware Boot instead of directly starting the OS kernel.
- For x86_64 hosts using the KVM hypervisor, the limit on the number of available virtual processors (vCPUs) has been increased from 254 to 8192.
- The virtio-blk device implements an optimization that uses grouping and batch processing of asynchronous I/O requests. The change improves performance with small block sizes (<=16KB).
- The execution of the pause operation has been accelerated virtual machines, which is especially noticeable in virtual machines with a large number of vCPUs.
- Added documentation on running guest systems under Cloud Hypervisor management. Windows 11.
- Support for Intel SGX (Software Guard Extensions) technology has been discontinued.
Source: opennet.ru
