Sample Admission Essay Paper on Server Virtualization


The information centric economy and the internet continue to change the way businesses perform their daily activities (McAllister, 2007). More business organizations, financial institutions and even educational institutions are relying more on the internet and information for service delivery and execution of daily activities. With this reliance, there has been a surge in the use of technological equipment and space for this equipment. Servers have specifically been important elements in the provision of internet services as well as resources required by different divisions in an organization. With increased resource use, especially for educational institutions, there is need for use of even more servers, which require more space, while at the same time increasing the cost for the institutions with regard to space, power and purchase of the new server hardware and personnel for monitoring and maintenance. Additionally, the continually dynamic and intricate nature of the IT world necessitates the need for better technologies to handle the resource and computing needs of organizations. Fortunately, however, the new technological use of server virtualization provides a solution not only optimize to space for organizations but also to save on the cost in power for cooling and purchase of new hardware. A number of virtualization productsare present in the market. These include VMware’s vSphere, Microsoft’s Hyper-V and Citrix’s Xen Server. In order to choose the best technology for an educational institution, it is important to evaluate each of the available technologies, making a comparison among them, on their Level of Availability, Fault tolerance, Amount of downtime Guest, Operating Systems supported, Hardware requirements, Security, Ease of use, Tools, Licensing and Service support for a decision on which technology best fits.

Server virtualization is the process of masking server resources from users using dedicated software, and dividing these resources into smaller virtual servers to help in the maximization of the server resources. While it is not a new concept given that mainframes in the 70s ran numerous processes of an operating system simultaneously with independence from the other processes the recent advances in software and hardware have made the process a standard, especially on commodity servers (McAllister, 2007). Data center managers therefore today have a wide array of virtualization solutions to choose from given the recent advances in virtualization technology.

While there are different products available in the market for visualization such as VMware’svSphere, Microsoft’s Hyper-V and Citrix’s Xen Server, all these are based on the three basic virtualization technologies, which are full virtualization, para-virtualization and OS centered virtualization, also referred to as OS partitioning (McAllister, 2007). The choice on which technology to use for best results depend on the workloads of each server and the priorities that each server will need to operate (McAllister, 2007).

The widely popular among the virtualization technologies is full virtualization. This technology uses a hypervisor, a software used to create an abstraction layer between virtual servers and the embedding hardware (McAllister, 2007). Both the Microsoft and VMware use full virtualization in their technology, while Citrix is a cloud based server product (Citrix, 2014). The technology’s hypervisor captures the CPU commands and facilitates access to hardware controllers and exterior resources, thus allowing the installation of an OS on the virtual server, without the need for making any alterations, and the server itself having no knowledge of its running in a virtual environment (McAllister, 2007). In essence, full virtualization allows the hypervisor to manage the virtual servers, therefore making the servers guest OSes. This process is processor intensive given the need for the hypervisor to not only manage the servers (virtual), but also to maintain their independence of each other.

Para-Virtualization presents a solution to the overburdened hypervisor, making the servers aware of their running in a virtual environment and therefore allowing them to coordinate and work together with the hypervisor (McAllister, 2007). Citrix’s XenServer, an open source server is an example of para-virtualization (Citrix, 2014). Para-virtualization requires the adjustment of the OS at the Kernel level for smooth running as a virtualized OS. For this matter therefore, para-virtualization is largely feasible for open source system, and not for proprietary systems such as Windows, that does not allow modifications. Given the coordinative and cooperative nature of the virtual servers and the hypervisor, there is increased performance with near unvirtualized server response (McAllister, 2007).

Operating system based virtualization on the other hand, denotes building virtualization capability at the OS level. This virtualization technology features the absence of a hypervisor, with the OS being responsible for the division of hardware resources among the numerous virtual servers, keeping the servers independent of each other (McAllister, 2007). Within such a system, all the virtual servers run a similar operating system, with each server having its own application and user accounts. The technology is however less flexible, but with better native-speed performance. Additionally, management of such a system is much easier than the heterogeneous environment present in full and para-virtualization systems (McAllister, 2007).

Given the existence of the three technologies, the choice on the technology to use is dependent on the demands of the organization. As a university, the nature of use of the servers is varied and therefore the need for a versatile technology to manage such demands effectively. Virtualization presents a number of benefits for such an institution, not only it saving money on the purchase of other server hardware, but also in space. One of the current and rapidly spreading up trends in the technology world is cloud computing. Among the merits of the cloud are availability, flexibility and cost saving. By virtualizing the servers, the institution will be making a step towards cloud computing and ultimately enjoying the benefits of cloud computing (Marshall, 2011).

For the institution, although the computing platforms are aged, some are still running legacy applications. These platforms, in addition to running the legacy applications, may not be open for updating. By virtualizing the university system and condensing these applications within the virtual environment, it is possible to extend the life of the application, avoid any outages and rid the institution of the old machines running the applications (Marshall, 2011). In this case, virtualization helps in the smooth transition from the old into the new system, without causing outages to users within the system.

Among the major goals of server virtualization is the reduction of the sum of physical servers. As a characteristic of virtualization, servers are partitioned and isolated, which in essence enables a simple and safer consolidation of servers (Metzler, 2011). Consolidation subsequently cuts the number of physical servers. Therefore, there is a huge reduction in server floor space, power utilization and air conditioning costs.

Given the varied divisions and faculties within a university, different divisions may be in need of different computing resources at a given time. By virtualizing, it is possible to make computing reallocations to different divisions and applications as need dictates within seconds or minutes without interrupting service to other applications or users (Metzler, 2011). Among the features present in virtualized servers are “live migration, storage migration, fault tolerance, high availability, and distributed resource scheduling” (Marshall, 2011). All these work to ensure that users get unparalleled and seamless services, while avoiding unplanned outages.

Server virtualization additionally helps in saving time for the institutions. Physical servers are time and personnel intensive from ordering, installation, configuration and testing. The processes of making a physical server reading for deployment could take weeks, while with a few clicks, a number of virtual servers could be set up and be ready to start working. Even with the needs of the institution growing in the future, a virtual machine could be easily copied and loaded with applications and be ready to working within minutes. This can however not be said of physical servers given the aforementioned process to have them running (Marshall, 2011). A similarly easy process is also executable when the institution does not require a virtual machine; it can easily be deleted from the main server.

The flexibility and security that virtual servers provide are phenomenal. Given the isolated nature of virtual machined from each other, a crash, virus attack, failure or compromise of one virtual machine does not affect the others, as well as the physical server. The failed server, its applications and data can moreover be restored within a short time and with a small number of clicks. Shutting down, backing up, replacing, and restoration of a virtual machine is much easier in comparison to a physical server that has been compromised (Principled Technologies, 2011).

Given such benefits that accrue in using virtualization, it is important to do a comparison of the three products available for the institution. While there are many other virtualization products, the three (VMware’s vSphere, Microsoft’s Hyper-V and Citrix’s Xen Server), are the products available for the institution, and therefore form the basis of comparison. All the three technologies run on the Bare Metal hypervisor type, giving them almost similar coordination and cooperation capabilities (Venezia, 2011). Additionally, all the three run hardware assisted virtualization, although some among them offer other virtualization typologies embedded within the same software offering.

For the architecture, they are all based on the x64 and x86 architecture, and can therefore run on systems with the two architectures, this is in addition to their capability on being deployed using the On Premise model, on both Windows and Linux platforms (Venezia, 2011). This gives all the three a wide range of models within which they can be deployed on systems running different platforms.

For their differences, the first and most noticeable is the pricing and licensing. Given that Citrix’s XenServer is an open source software, it is free of charge while Microsoft’s Hyper-V and VMware’s vSphere have different pricing and licensing options. For Microsoft, its software has an $882 tag for the processor and the CAL (Venezia, 2011). On the other hand, VMware’s software is priced at $2,875 for the vSphere 5 enterprise that includes one processor, which has a 64 GB vRAM entitlement per processor (Venezia, 2011). Both Microsoft and VMware’s software have proprietary licensing, although Citrix offers per-server licensing, while VMware offers per processor licensing. Microsoft conversely has no licensing details.

Citrix’s XenServer targets both personal and small-to-medium sized businesses. In contrast, Microsoft’s Hyper-V server targets the small-medium businesses, while VMware’s vSphere targets enterprise customers (Venezia, 2011). In essence, therefore, Citrix has a larger offering base (personal and small-medium business) than the two other contenders do.

The range of storage supported by Citrix is also large given that it supports USB, NFS and NAS storage. In contrast, of the popular storage formats, VMware supports SSD for Swap, which both Citrix and Microsoft do not supports. Moreover, Microsoft does not support any of these popular storage formats (Venezia, 2011).

The three additionally offer a wide range of virtual machine limits in virtual disk size, RAM per virtual machine and virtual CPU per virtual machine. For virtual disk size, Microsoft tops the list with an estimated 2040 GB per virtual machine. Both Citrix and VMware tie at 2000 GB of virtual disk size that they support (Venezia, 2011). On the other hand, VMware has a higher RAM per virtual machine at 1000 GB, followed by a Citrix at 128 GB while Microsoft trails the list with only 64 GB of RAM. VMware further has 32 CPUs per virtual machine, with Citrix having 16 and Microsoft trailing with four (Venezia, 2011).The amount of virtual RAM and CPU per virtual machine allows VMware to host up to 604 virtual machines per host, while Microsoft and Citrix host 384 and 75 respectively. Conclusively, Citrix allows for both online and on premise deployment and platforms. The other two however, only support on premise deployment and platforms such as Windows and Linux.

Each of the three offerings offers an assortment of features for their customers. Top among these features include availability, fault tolerance, downtime time and guest operating systems supported. While each may have these features embedded, the functionalities of these features may be different and therefore affect performance.  VMware has a simple high availability configuration embedded within the virtual center. On the other hand, Microsoft’s high availability is configured through setting up of the Windows Fallover Clustering. For Citrix, the configuration requires setting up of power management upon each host, as well as per virtual machine, making the set up process not only cumbersome but complicated (VMware, 2014).

VMware offers live migration capabilities, a feature only currently present on the vSphere (Venezia, 2011). This minimizes downtime and allows continuity in service even during migration from one system to another. Such a feature therefore makes VMware the only virtualization technology that can ensure few downtimes even during peak operations without compromise to the users.

The licensing and user support for the three is different. Both vSphere and Hyper-v have complex licensing than XebServer’s. VMware offers licensing over its wide range of products with different levels, with added features for a more priced tier per physical socket. For Microsoft, the licensing is attached to Windows Server. Enterprise licensing allows for four virtual servers on a physical server with a similar operating system, while a datacenter license allows for unrestricted number of virtual machines for every server (Venezia, 2011). Citrix has license per server with no restriction on the server capacity and offers a choice of several bundles of the server license.

All the three however offer some sort of technical support to customers. Microsoft specifically has 24-hour tech support for its customers in addition to help desk, webinars, recorded demos and request forms. From these, customers can find support for any problems that they face with the product. On the other hand, Citrix offers FAQ, instructional videos and internet-based self-help. While these provide helpful information, they are limited as none gives a one-on-one technical assistance to the customer. For VMware, tech support includes FAQ, helpdesk, mailing list, onsite support, online self-service and webinars. These in essence, offer more support to customers, especially the onsite, help desk and webinars.

VMware and Citrix both bundle virtual host and virtual machine management within the same server making both of them easy to use. These tools therefore allow datacenter managers to easily monitor, create, delete and even update virtual machines from within one management server (Venezia, 2011). For Microsoft, host and virtual machines management tools are tasks spread across several system control tools. To get such advanced tools, users pay more, therefore adding the overhead cost, set up intricacies and complexities for the data center managers (Venezia, 2011).

Microsoft’s complexity issues also extend to number of guest operating systems that it can support. The number of supported guest operating systems is spread thin for Microsoft being able to support Fedora, CentOS, Ubuntu, Windows 7, Windows Server 2008 and Vista. Both Citrix and VMware support a substantial amount of guest OSes, with VMware supporting almost all the guest OSes. This makes the two much more adaptable than Microsoft’s Hyper-V.

For the host requirements, VMware’s vSphere has the lowest in minimal requirements with a wider range of processor compatibility. At a minimum speed requirement of 400MHz, it offers the lowest requirement in processor speed, with the ability of running on Celeron, Pentium II-4, as well as a wide range of AMD’s processors. This is in comparison to the 1.4 and 1.5 GHz required for Citrix and Microsoft respectively. Additionally, vSphere has a lower RAM requirement at 128 MB, while the others both require a minimum of 1 GB of RAM.

VMware provides more offing in support, features, availability and level of tolerance than any of the other technologies. Additionally, VMware’s vSphere supports more guest operating systems, more management features, more virtual machines per host and a large virtual disk size. Given the resource intensive nature of a University therefore, and the number of virtual machines that the university may require, VMware’s vSphere is the best candidate for the university as it will not only provide versatility, but also speed and great performance to the institution.


Citrix.(2014). Citrix XenServer.Citrix. Retrieved from

Marshal, E. (2011).10 Benefits of Server Virtualization. InfoWorld. Retrieved from,2

McAllister, N. (2007). Server Virtualization.InfoWorld, 29(7):20-22

Metzler, J. (2011). Virtualization Benefits: White Paper.Riverbed Technologies.Retrieved from

Principled Technologies (2011).Virtualization Benefits.Principled Technologies. Retrieved from

Venezia, P. (2011). Virtualization shoot-out: Citrix, Microsoft, Red Hat, and VMware.InfoWorld.Retrieved from

VMware.(2014). Comparison Guide.VMware. Retrieved from