The Vladimir state university made a choice for benefit of Hyper-V

The Vladimir state university — the largest state university in the region. On its balance more than 2500 PCs and over 50 servers consist. At the university a large number of basic and applied applications and services which users are students, the staff of research divisions, an administrative and management personnel functions. The university dynamically develops, so, also his requirements in the field of IT constantly grow. Among tasks, important for the university — consolidation of IT infrastructure, increase in its reliability and efficiency of use of resources, cutting of costs for service of IT assets.

In the beginning in university the pilot project which proceeded about four months was executed. In a test environment the virtualization platform from Microsoft was deployed. As a reference model the standard architecture of Windows Server Reference Architecture (WSSRA) which fragments were adapted for requirements of university was taken. After studying, setup and working off of the main opportunities, the decision on implementation of virtualization in the IT environment of the university was made.

Situation

The Vladimir state university — the largest state university in the region. On its balance more than 2500 PCs and over 50 servers consist. At the university a large number of basic and applied applications and services which users are students, the staff of research divisions, an administrative and management personnel functions. The university dynamically develops, so, also his requirements in the field of IT constantly grow. Among tasks, important for the university — consolidation of IT infrastructure, increase in its reliability and efficiency of use of resources, cutting of costs for service of IT assets.

In November, 2009 it became obvious that means for the solution of these tasks are technologies of server virtualization which allow to save significantly funds both for maintenance, and for gradual deployment of services, and give more flexible opportunities for management of the IT environment and its further scaling. At that time in university over 30 servers were used. Most of them were insufficiently reliable and productive, obsolete. Therefore it was necessary to replace both a hardware component of DPC, and the platform of management.

The university considered two virtualization platforms: Microsoft Hyper-V and VMware.

Advantages of Hyper-V became simplicity of implementation and operation (deployment and support of VMware required additional training), as a result — higher speed of implementation. All critical applications at the university worked under operating systems and Microsoft services therefore Hyper-V optimized for support of the "native" environment was more preferable and reliable as for both deployment of virtual machines, and migration from a physical medium in virtual, and supports of stable full functioning. Besides, Microsoft offered the developed infrastructure monitoring system — a packet of System Center. And if the VMware platform managed only virtual environment, then Microsoft System Center integrated control of physical and virtual servers and a client part in the single console. Also the fact that Microsoft offers universities of the license according to the special Software Assurance Academic program allowing to optimize costs and to distribute them for several years became important.

For these reasons the main virtualization platform from university selected Hyper-V. Nevertheless, at the university also the system on the basis of VMware of ESXi licensed for the academic use is used. Both of these platforms were implemented at the same time that allowed IT specialists to compare in practice of a possibility of these products.

Solution

In the beginning in university the pilot project which proceeded about four months was executed. In a test environment the virtualization platform from Microsoft was deployed. As a reference model the standard architecture of Windows Server Reference Architecture (WSSRA) which fragments were adapted for requirements of university was taken. After studying, setup and working off of the main opportunities, the decision on implementation of virtualization in the IT environment of the university was made.

The large-scale implementation project of technologies of virtualization began with purchase and installation of more powerful and modern server hardware of IBM and HP. On its basis two DPCs were created, each of which was focused on the solution of the tasks. The first DPC is expected support of office-work and scientific activity. It is unrolled on 18 edges of IBM BladeCenter and 10 HP servers of Proliant from which about 20 physical servers are virtualized based on Hyper-V, including eight servers of IBM are organized in a format of a failover cluster that allows to involve CSV and Live Migration technologies. In the second DPC there are two separate physical servers, also under control of Hyper-V on which Active Directory services, critical for the organization, with six controllers of domains are started.

4 physical servers, now them already 12 were given for virtualization of applications initially. Applications were transferred to virtual environment consistently, one behind another — Lotus, Parus, the library system, infrastructure of some scientific projects (the websites, application servers), the high school portal on the Sharepoint platform, some products of the Forefront family. Migration of the mail Exchange system comes to the end. Migration happened so smoothly that users did not feel any discomfort and only noted the increased performance and big stability of operation of applications.

At implementation of virtualization means of P2V and also the products System Center — Operations Manager, Configuration Manager, Virtual Machine Manager, Data Protection Manager were used what significantly reduced costs for migration. Later, with a release of the version of Hyper-V 2008 R2 SP1, there was an opportunity to use runtime allocation of memory, i.e. to increase even more efficiency of the equipment.

Integration by means of Virtual Machine Manager (VMM) allows to control the events on both platforms, Hyper-V and VMware. VMM uses the API interface of the management console Virtual Infrastructure Manager from VMware and thanks to it gets an information access about servers of VMware. Thus, the administrator receives a uniform picture about all virtual environment. The Performance and Resource Optimization (PRO) technology entering into VMM activates mechanisms of observation of virtual machines, and in case of excessive load of the physical server a part of virtual machines are automatically transferred to other servers. Thus, the fault tolerance and performance of DPC increases.

In addition to VMM intended for the organization of virtual infrastructure, Configuration Manager (SCCM) is focused on management of physical IT assets — more than 50 servers and more than 2500 PCs. SCCM is responsible for functioning of the hardware, software, installation of updates, etc. Among other tasks, SCCM helped to prepare network for implementation of technology of protection of access of Network Access Protection.

Operations Manager (SCOM) allows to control the main characteristics of the physical servers working under management both Windows and Linux. In case of incidents with the equipment or the applied systems, SCOM generates notifications to IT administrators. Via the web console control can be exercised far off. Reservation of servers, physical and virtual, is executed by means of Data Protection Manager (DPM).

Advantages

"Virtualization allowed to increase the general reliability of systems and services, to save due to consolidation on the server hardware of high performance (including taking into account expenses on the electric power and conditioning) and also, thanks to reduction of labor costs on administration and development. As we use at the same time two virtualization platforms, Hyper-V and VMware, we have an opportunity to compare their convenience, functionality and efficiency. In our case Hyper-V wins on a number of factors, such as centralization, functionality, total cost of ownership. And products of the System Center family do the constructed heterogeneous system flexible and managed and also allow to perform the pro-active analysis of deniability with automatic reaction to an incident" — the director of the Regional center of new information technologies of VLGU Sergey Roshchin claims.

The highest reliability of the solution is provided due to use of failsafe configurations of the equipment at the physical layer and at the level of OS. Time of failure is almost reduced to zero. IT infrastructure of DPC adapts under the growing requirements of university. At the same time all data are reliably protected from loss thanks to dynamic archiving. Actually the idea of providing resources on demand — resources, such as RAM or processor is realized in DPC, are increased quickly and without problems, but not used virtual machines can be removed in archive.

For infrastructure services cost is significantly reduced both the equipment, and ownership – most services are consolidated on a small amount of modern physical machines with a large supply on scaling and reliability.

At the university it is created and the library of images allowing even to technically not prepared users to create quickly virtual machines with the set characteristics is constantly updated. So, virtual machines for administrative divisions and scientific groups are controlled by the staff and graduate students of departments through a self-service portal. The virtual machines used for training of students are created by teachers of rates before occupations or students directly on occupations. Thus, the possibility of extremely flexible and quick access of students to different software for studying, passing examinations is reached. Use not of physical, but virtual environment protects computers in educational classes from accidental damage of the operating system. Besides, on the same PCs without double loading and reconfiguration it is possible to use Windows platforms and Linux, it comprehensively training promotes training of high-class specialists.

Instead of 5-6 administrators, now IT infrastructure two persons manage to drive. Many tasks do not require departure in the remote body any more, and are solved far off. Being at home or in a business trip, via the Internet administrators can enter on any server irrespective of it is included or is switched off, to check its work, far off to set software. As the university is located in several body and has two branches in other cities, it considerably reduces troubleshooting terms. Earlier in DPC there were 12 servers, each of which consumed on 300-400 W. Now resources are consolidated by four machines, each of them consumes at the moments of peak loads on 200-300 W. Thus, taking into account technologies of dynamic energy management of servers, cost reduction due to energy saving is at least 4-5 times.

By rough estimates, the cost of ownership of DPC fell more than by 10 times due to reduction of quantity of components which can refuse and also due to implementation of technologies of a failover cluster and migration online. From the moment of implementation of virtualization at the end of 2009 unplanned idle times of services did not happen.