‘Virtualization’ in information technology parlance is the process of breaking a single physical resource into multiple logical / virtual resources. This can be better understood through an example of server virtualization.
An organisation ‘X Ltd.’ has a high end IBM server with ‘Windows Server 2008’ operating system. On this server, X ltd has been running an enterprise application developed with Microsoft tools. The server’s resource utilization (CPU, RAM and Disk utilization) has been below 40% of the prescribed peak capacity. The company now has plans of implementing an open source email application on Linux platform. In a pre-virtualization environment, the IT department would have bought an additional server to install the Linux OS in order to run the email application. This would have resulted in additional investment spend, extra hardware maintenance, additional space & power and additional efforts in terms of configuring controls in the new server. All these can be avoided through virtualization.
The IT department of X Ltd has decided to virtualize their existing IBM server. This means this IBM server will have two OS viz. Windows Server 2008 and Linux. How can this be achieved? X Ltd. will have to install a virtualization software (broadly termed as ‘hypervisor’ or ‘virtual machine monitor’ (VMM). This hypervisor software will sit on top of the host operating system (windows server 2008). Onto this hypervisor, the Linux OS will be installed. The email application will then be installed in the Linux OS environment. Through this process the CPU, RAM, network card and disk resources have been shared between two different environments either dynamically or through a pre-configured method. This way resources can be optimally used. This kind of virtualization is ‘Hosted’ virtualization as the hypervisor is hosted on an existing OS. There is another kind of virtualization where there is no OS on the hardware and hypervisor sits directly on top of the hardware as hardware control and guest OS monitor. Such kind of hypervisors are called bare-metal or native hypervisors. In this scenario, if X ltd had gone in for the native hypervisor, the Windows Server 2008 and Linux will be sitting on top of the hypervisor while the hypervisor acts as the intermediary between these OS and the IBM server.
There are many scenarios where server virtualization is helpful and beneficial. One in case of expansion as mentioned above. In another scenario, an application may be running on a piece of hardware that has seen end-of-life and the OS is different from the OS in other servers. Then this legacy application and related OS can be moved to another server through virtualization and thereby the old hardware can be retired. Virtualization is also useful for testing environments whereby a portion of server can be dedicated for testing new things rather than investing in new servers.
Comments are closed.