Microsoft Hyper-V

Microsoft Hyper-V, the code name Viridian, technology earlier known as Virtualization of Windows Server (Windows Server Virtualization) — the system of virtualization on the basis of a hypervisor for x64 systems. The beta of Hyper-V was included in x64 version of Windows Server 2008, and the final version for these versions was released on June 26, 2008.

The free separate version of Hyper-V which received the name "Microsoft Hyper-V Server 2008" was released on October 1, 2008. Is the basic ("Server Core") Windows Server 2008 option, i.e. includes all functionality of Hyper-V; other roles of Windows 2008 Server are disconnected, there are also restrictions of Windows services. [4] Free 64-bit "Core" - the version of Hyper-V is limited to the command line interface (CLI) where the configuration of the current OS, the physical hardware and program equipment is executed by means of cover commands. The new menu of a control interface executes simple primary configuration, and some freely extended scripts expand this concept. Administration and configuring of the virtual server (or guest OS) is performed by means of software installed on the PC running Windows Vista or Windows 2008 Server with the set addition for administration of Hyper-V from MMC. Other option of administration / configuring of the North Windows 2008 Core is use deleted with Windows Desktop or Server at redirection (some) Management console (MMC) ukazyvyushchy on Core Server. It considerably simplifies setup, reducing it to several "cliques" of a mouse.

The private cloud of Hyper-V Cloud constructed based on the products Microsoft Windows Server and System Center allows to lower considerably operating costs of support of a core infrastructure, to automate service and to reduce risks of idle time of IT systems. Implementation of Hyper-V Cloud simplifies providing IT as service, allows to consider real contribution IT to business of the company. In further Hyper-V Cloud will become a reliable basis for creation of flexible hybrid model of use of cloud computing – placements of IT services or their components in a cloud for choice: private, service provider or public Microsoft Azure. Type selection of a cloud depends on business needs of the company.

At the Build conference which took place in September, 2011 the Microsoft company disclosed some details about new opportunities of the updated virtualization platform based on Hyper-V 3.0 as a part of server Windows Server 8 OS. In Hyper-V 3.0 the set of interesting opportunities in the field of server virtualization and VDI infrastructure will appear:

• Support up to 160 logical processors on Hyper-V hosts
• Support up to 2 TB RAM a host server
• Support up to 32 virtual processors of virtual machines (vCPU) and up to 512 GB of RAM on one VM
• Support of NUMA technology in guest OS that allows to increase the capacity of the virtual machine
• Support of several simultaneous "hot" migrations (Live Migration)
• Support of the Storage Live Migration equipment - migration of storage of the virtual machine without interruption of its work (it will work without the need for existence of the general storage - at local disks at the expense of the built-in opportunities of replication)
• New format of virtual disks of VHDX which allows to create virtual disks up to 16 TB (instead of 2 TB in the VHD format). Also new format will be more productive, reliable and to support work with big blocks
• Offloaded Data Transfer (ODX) technology which allows to transfer to the party of the disk array of operation on work with storages of virtual machines (at VMware the similar technology is called vStorage API for Array Integration, VAAI)
• The updated virtual switchboard giving enhanced capabilities on virtualization of networked environment virtual machines and a possibility of its portability between cloud infrastructures (something similar to VXLAN technology from VMware and Cisco)
• Support of virtual adapters (Virtual Fibre Channel) for virtual machines (to 4) through which VM can get direct access to LUN by means of Multi-Path I/O (MPIO). Also there will be virtual Fibre Channel-switches
• Support of loading of hosts through Fiber channel and iSCSI SAN.
• Support of SR-IOW technology for exclusive access to PCI devices
• Expanded monitors of performance (CPU Metering and another)
• Pools of resources for clusters (Resource pools)
• Support of the deduplication of data allowing to reduce tebuyemy virtual machines space on storage system without significant performance penalty. It will allow to reduce backup windows also
• Direct transmission of data between hosts at the expense of Offloaded Data Transfer
• Support of NIC Teaming and load balancing for creation of virtual networks on a host (earlier it became using third-party drivers)
• Support of arrays of JBOD and thin disks on JBOD (Thin Provision)
• Support of means of Bitlocker for cluster disks
• The new version of the Cluster Shared Volume (CSV) 2.0 file system with the built-in support of deduplication and creation of snapshots from arrays
• Means in GUI for management of the IP addresses (IPAM)
• Support of CIFS/SMB storages using the Remote Direct Memory Access (RDMA) protocol
• The Hyper-V Replica technology allowing to organize asynchronous replication of virtual machines (we will remind that in Veeam Backup and Replication 6 there will also be such opportunity)
• Support of RemoteFX technology for a RDP session with a host, the improved traffic compression for WAN channels
• Possibility of creation of a basic image for virtual PCs (gold master image). Individual sessions of users can be configured using the moved profiles (roaming profiles)
• Support to the 63rd a host servers and up to 4000 virtual machines in a cluster
• Support of graphic libraries DirectX 10, OpenGL 1.1 and Metro UI technology in virtual machines
• Services Active Directory will be finished under virtualization (support of snapshots VM, virtualization of domain controllers and their cloning)
• Built-in broker of connections with a possibility of balancing of sessions
• Possibility of inclusion and shutdown of the GUI server, turning it into Server Core (and back)
• Support of the distributed virtual switchboard Cisco NEXUS 1000V which will be specially developed under Windows Server 8

Microsoft Hyper-V 3.0

With a release of Windows Server 2012 also the system of virtualization - Hyper-V was updated. In this article is considered what new opportunities appeared in Hyper-V 3.0. It should be noted that the third version contains many improvements of functionality in comparison with the previous version, and, the main thing, - for the first time in the history of Hyper-V, Microsoft reached the level of products of Enterprise-level, comparable to VMware vSphere 5.

The main management tool and monitoring of Hyper-V is Hyper-V Manager. Starting with version 3.0, the set of new features of management became available through PowerShell. For example, start of the VM1 virtual machine can be executed using the Start-VM vm1 command. Further key features of the new version of Hyper-V will be considered.

Memory Ballooning

Technology of memory allocation in such a way that the total extent of the memory selected to separate virtual machines of the server of virtualization exceeds amount of the memory which in general is available on this server of virtualization is called Memory Overcomitment, and its development in Hyper-V 3.0 - Memory Ballooning. These technologies allow to increase coefficient of consolidation of servers through increase in density of placement of virtual machines (the number of virtual machines at the rate on the virtualization server). This result is achieved due to dynamic memory relocation, selected to separate virtual machines, at the same time the activity of the virtual machine is considered, its status it (is included, switched off), such parameters as "the minimum amount of memory" and "amount of memory at a stage of start of the virtual machine" (the administrator can determine them). And, if in Hyper-V 2.0 these two parameters were actually equivalent, then in Hyper-V 3.0 they completely independent, - for example, for a phase of start of the machine it is possible to select more memory, than that minimum which will be received by the virtual machine in a work phase, but the failure to act period.

Live Migrations

Live Migration (migration of virtual machines between virtualization servers without the need for their switching off) appeared in Hyper-V 2.0. Initially there was a restriction - it was possible to move only one virtual machine of the server of virtualization at the same time. In Hyper-V 3.0 this limit is lifted - several virtual machines it is possible to migrate in parallel, at the same time the administrator can set restriction for the number of at the same time moved virtual machines to avoid an overload of network.

Earlier migration of virtual machines within Live Migration was possible only within cluster nodes. In Hyper-V 3.0 of this restriction does not exist any more. Such "transcluster" movement of the virtual machine is called Shared Nothing Live Migrations. It should be noted that this possibility should not be considered as long-term replacement to traditional clusters, and should be used mainly only for actually movement of virtual machines of the required server of virtualization.

Requirements for Shared Nothing Live Migration:

• At least two servers of virtualization Hyper-V 3.0 Servers should be included in one domain Active Directory For ensuring maximum capacity of the server should have processors with identical architecture (see the description of compatible processors in documentation on Hyper-V 3.0) Servers should be connected by gigabit or faster network connection of Virtual switches should have an identical configuration Client for Microsoft Networks and File and Print Sharing for Microsoft Networks Should be included
• The general storage (shared storage) is not necessary for Live Migration work any more as it was to Hyper-V 3.0. Now local storages can be used, and data migration will be executed on Ethernet. The same new technology allows to build clusters also without the general storage now.

Storage Migration

This technology differs from Live Migration in the fact that files of the virtual machine from one storage move to another while the virtual machine continues to be performed on the same server of virtualization. This technology replaced Quick Storage Migration technology, being available in Hyper-V 2.0. Happens in Hyper-V 3.0 storage migration of the virtual machine without switching off or suspension of its work (unlike the previous version of Hyper-V).

Hyper-V Replica

This technology allows to replicate the working virtual machine on other server of virtualization. Ikrementalny changes are transferred on the configured schedule with the minimum interval of start, close to real time in terms of backup ("Near CDP") - approximately to 5-15 minutes. Hyper-V Replica can work at lines with high time delays and unstable communication. This technology are positioned Microsoft as means of protecting from catastrophic failures.