The Xen open source virtual machine monitor was originally created through the University of Cambridge Computer Laboratory and developed through XenSource, Inc. Citrix Systems acquired XenSource in 2007, and Xen technology has since emerged in the free edition called XenServer, along with two paid enterprise-class products: Essentials for XenServer Enterprise and Essentials for XenServer Platinum. Other commercial implementations of Xen include Oracle VM and Sun xVM.
The Xen open source virtual machine monitor is designed for common Intel and IBM architectures, supporting x86 (32 and 64-bit), Itanium, and PowerPC-based systems. The Xen hypervisor loads and supports all of the subsequent operating systems (OSes) and workloads. This is referred to as a Type 1 or bare-metal hypervisor, which runs directly on the system's hardware and hosts OSes above it. The first guest is typically the system's host OS, which has management privileges and extensive control over the system's physical hardware (Xen nomenclature denotes this "domain 0" or "dom0").
This allows system administrators to access and control the server and other workloads from that first guest virtual machine (VM). Xen management tasks are usually simplified and automated through third-party tools like XenTools, Ganeti, MLN, HyperVM, Convirtue and others.
There are many operating systems that can serve as a host operating system (dom0). These include most distributions of Linux, Novell's SUSE Linux Enterprise Server release 10, Red Hat Enterprise Linux 5, Fedora, openSUSE 10.3, Ubuntu 8.04, NetBSD 3.x, Debian release 5 and others. Xen 3.0 and later can support Microsoft Windows and other proprietary OSes as guests, but the host server will require processors that support virtualization acceleration technologies such as Intel VT and AMD-V.
The hypervisor in the Xen open source virtual machine monitor is particularly noted for its virtual machine live migration capabilities, allowing administrators to move virtual workloads from one physical host to another without shutting down or even disrupting the workload being moved. Generally speaking, a live migration process copies the memory space used by the VM and then replicates that content on the destination server. Once the copy is complete, the two iterations are synchronized and processing is handed to the destination server, allowing the original server to delete the unneeded instance. The actual disruption involved in this handoff is usually less than half a second.
Live migration is one of the key benefits of server virtualization because it enables several critical benefits related to application availability. The workloads on a problematic server can often be migrated (failed over) to other servers before a crash actually occurs, and this preserves the workload's availability. The ability to move workloads also makes it possible for administrators to balance the computing demands on each server. This optimizes system performance and helps to prevent VM (or even entire server) crashes due to inadequate computing resources. Finally, an administrator can migrate workloads to free up a server for maintenance or replacement without interrupting the workload's availability.