Server Virtualization: A Method to Maximize Return on Investment Definition of Server Virtualization
Server virtualization is an environment in which there is one physical server and there is software running, called a hypervisor, which allows several independent instances of operating systems (i.e. virtual machines) to run simultaneously. Ou (2006) states the server virtualization is the best solution for “applications that are meant for small- to medium-scale usage.” If implemented properly, organizations can use server virtualization to maximize their return on investment (ROI) in server hardware. Energy costs are rising and replacing “real estate as the primary data center expense” (Overby, 2007). Therefore, utilizing server virtualization represents a significant cost savings for organizations. The following three basic technologies are used to implement server virtualization: full virtualization, OS partitioning, and paravirtualization. (see Figure 1) Full virtualization uses the hypervisor to fully imitate all of the hardware resources of a physical server. The operating systems installed in a fully virtualized environment function as if they are the only operating system running. In other words, they are not aware of the other installed operating system instances running on the server. Therefore, a major advantage of full virtualization is flexibility. Different operating system versions and patch levels can be installed on the same server. The major disadvantage of full virtualization is performance. Because of processor overhead, server performance decreases. “Full virtualization is processor-intensive because of the demands placed on the hypervisor to manage the various virtual servers and keep them independent of one another” (McAllister, 2007). OS partitioning, also known as Single Kernel Image (SKI), integrates the hypervisor with the operating system. The major advantages of OS partitioning are performance and manageability. OS partitioning is not processor-intensive like full virtualization; therefore, overhead is reduced, which in turn increases server performance. Furthermore, OS partitioning is easier to manage, because each installed operating system instance must be the same. The major disadvantages of OS partitioning are flexibility, reliability, and security. As mentioned previously, different operating system versions and patch levels cannot be installed on the same server, and according to the article entitled, What is Virtualization? (2007), “if the kernel is exploited, all OS instances resident on the system will be compromised”. Paravirtualization, also known as hardware-assisted virtualization, uses a hypervisor like full virtualization, where a layer of abstraction is inserted between the physical server hardware and the multiple operating system instances. However, rather than emulating the hardware resources of a server, each installed operating system instance is modified to allow direct access to the hardware resources. Unlike full virtualization, the operating systems installed in a paravirtualized environment are aware that they are running in a virtual environment. Both of these differences reduce processor overhead. Similar to full virtualization, different operating system versions and patch levels can be installed on the same server. Therefore, in addition to the flexibility offered by full virtualization, paravirtualization also offers an increase in server performance similar to OS partitioning. Unfortunately, paravirtualization technology is not yet available for Windows systems, because the required modification of the operating system is at the kernel level, and proprietary systems, like Windows, cannot be modified (McAllister, 2007). Figure 1. Three Faces of Server Virtualization (McAllister, 2007).
Perot Systems Corporation is a Fortune 1000 corporation headquartered in Plano, Texas, with operations and offices worldwide,...
Please join StudyMode to read the full document