Release: Xen Project Hypervisor 4.5

The Xen Project, the community which develops the Xen hypervisor under the GNU General Public License (GPLv2) last week announced the availability version 4.5 of the Xen Hypervisor. With major contributions from AMD, Bitdefender, Cavium, Citrix, Fujitsu, GlobalLogic, Intel, Oracle, as well as several individual and academic institutions the product has been enhanced with several new features and capabilties.

The following new capabilities and features are available in version 4.5.:

  • Xen PVH virtualization, which is an extension to the classic Xen Project Paravirtualization feature which uses hardware virtualization extensions available in modern servers. PVH boots as the first guest and takes on the responsibilities of the initial domain known as dom0. This means Xen Project Hypervisor is able to take advantage of contemporary hardware features like virtual machine extensions (VMX) to significantly expedite execution of the initial domain. Instead of asking the hypervisor to handle certain operations, the dom0 can execute operations natively without compromising security. Additionally, improvements to the interrupt delivery mechanism for PCI passthrough workloads will help decrease latency and increase guest.
  • Support for Intel Resource Director Technology, for VMs running on top of Intel Cache Monitoring Technology (CMT) in order to monitor Last Level Cache (LLC) usage by application threads.
  • Coarse-grained Lock-stepping (COLO) allowing the state of a VM to be replicated on demand to a secondary VM on a different physical system.
  • ARM architecture updates, supporting larger VMs on ARM (up to 1TB of quest RAM), support for super page mappings, faster interrupt EOIs. Support for priorities and irq migration. Boot using EUFI firmware, now offering near feature parity with x86.
  • Introspection of HVM Guests Security, for protection against kernel exploits, zero days, rootkits and other malware attacks
  • New experimental multi-core anbled real-time scheduler allowing prediction of timing and performance of a VM
  • Support for Systemd

More details can be found in the following blog post.