Virtualization is not a new technology. It is an old technology harnessing the power of new technology (the Cloud) to deliver state of the art services to end users. Virtual machines come in three flavors—the hypervisor, host based virtualization and paravirtualization. Each of these VMs can be distinguished from each other by their positioning on the host machine.
Virtualization at the hardware level would require the use of Hypervisors. The hypervisor is also called Virtual Machine Monitor (VMM). It has the capacity to convert physical devices into virtual machines. The hypervisor may have micro-kernel architecture or a monolithic architecture. The hypervisor can install bare metal on the hardware and interact directly with it. An operating system may be installed on the hypervisor or the hypervisor may be installed on an operating system. Guest VMs may be installed on the hypervisor or underlying operating system and configured with operating systems of their own. Hypervisors with Xen architecture are open source hypervisor programs. This program was developed by University of Cambridge.
Host based virtualization installs the virtualization layer on top of the operating system. The host operating system remains in charge and manages the hardware for the guest systems that run on the virtualization layer. The host provides the device drivers and other low level services that are required and simplifies the VM design and deployment. While the host based virtualization architecture may slow down performance, it is flexible.
Paravirtualization is a process that modifies the guest operating system by providing specialized tools and application process interfaces for user applications. Performance may degrade in the process, portability, and compatibility may be in question and there may be variations in performance due to workload variations. However, paravirtualization allows virtualization to be introduced in different layers of the system so that virtualization overhead is reduced and performance can be improved by introducing virtualization only in specified guest kernels. The Guest OS will then be assisted by an intelligent complier that replaces the non-virtualizable OS instructions with hyper calls.
If you are aiming to transition to the Cloud post virtualization, it may be a good idea to examine the different virtualization options that are available to you before taking the first step in that direction. If you have already selected a Cloud vendor or are looking at a few vendors in a short list, you may like to discuss with them your options and see how your virtualization plans will fit in with your ultimate goal of “doing the Cloud”.
Securstore is happy to sit down with you and discuss your virtualization needs. Please contact us at any time.