Windows Server 2008 is designed from the ground up to support virtualization. This means that you have the opportunity to change the way you manage servers and services. With the Windows Server 2008 hypervisor, Hyper-V, there is little difference between a machine running physically on a system and a machine running in a virtual instance. That’s because the hypervisor does the same thing as a physical installation would by exposing hardware to VMs. The real difference between a physical installation and a VM running on the hypervisor is access to system resources.
That’s why we propose the following:
• The only installation that should be physical is the hypervisor or the Windows Server Hyper-V role. Everything else should be virtualized.
• Instead of debating whether service offerings—the services that interact with end users—should be physical versus virtual installations, make all of these installations virtual.
• The only installation that is not a VM is the host server installation. It is easy to keep track of this one installation being different.
• It takes about 20 minutes to provision a VM-based new server installation, which is much shorter than that of a physical installation.
• Creating a source VM is easier than creating a source physical installation because you only have to copy the files that make up the VM.
• The difference between a traditional “physical” installation and a virtual installation is the amount of resources you provide the VM running on top of the hypervisor.
• All backups are the same—each machine is just a collection of files, after all. In addition, you can take advantage of the Volume Shadow Copy Service to protect each VM.
• All service-offering operations are the same because each machine is a VM.
• Because all machines are virtual, they are transportable and can easily be moved from one host to another.
• Because VMs are based on a set of files, you can replicate them to other servers, providing a quick and easy means of recovery in the event of a disaster.
• You can segregate the physical and virtual environments, giving them different security contexts and making sure they are protected at all times.
• You can monitor each instance of your “physical” installations, and if you see that it is not using all of the resources you’ve allocated to it, you can quickly recalibrate it and make better use of your physical resources.
• Every single new feature can be tested in VMs in a lab before it is put into production. If the quality assurance process is detailed enough, you can even move the lab’s VM into production instead of rebuilding the service altogether.
• You are running the ultimate virtual datacenter, because all systems are virtual and host systems are nothing but resource pools.
Source of Information : McGraw Hill Microsoft Windows Server 2008 The Complete Reference