Understanding the Virtual Infrastructure Model
There is one piece of hardware supporting a single OS/Application instance in the traditional server architecture. As an example, this could be a corporate email server running Windows 2012/Microsoft Exchange. On average, an application like Exchange uses 15% of the processing capacity of the server hardware, leaving 85% of the processing capacity unused. Virtualization helps us address this issue.
In the Virtual architecture, a thin layer of software from VMware (called a Hypervisor) is inserted between the hardware and the operating system. The Hypervisor allows multiple OS/application servers (also called Virtual Machines or VMs) to exist on that same piece of physical hardware. This allows much more of the processing capacity of the hardware to be used to increase efficiency and resource utilization. Obviously, this would require less hardware in the data center to support the same number of application servers. The end result would be less capital spent on servers and supporting hardware.
To support this statement, let’s show an example of a bank that needs 12 application servers to support its operation. In the traditional model, the Bank would purchase 12 physical systems averaging $6,800 each (including hardware, operating system, and supporting hardware), with the total investment being $81,600. Assuming an adequately configured virtual server could support four application servers, the Bank would purchase three systems at approximately $9,000 each totaling $27,000. Then, the Bank would need to buy the operating systems and VMware software estimated at $25,000 and would probably want to purchase shared storage to take advantage of other benefits of virtualization. The cost of the shared storage and supporting hardware would be approximately $19,000.
Hardware $27,000 + Operating Systems
& VMware $25,000 + Storage $19,000 = $71,000
The virtual server infrastructure would cost $10,000 or 14% less than the physical server environment. And, with more servers in the data center, the percentage of capital savings can increase significantly. This is merely an example of how virtualization can save money in a data center. You can insert your own numbers, but the end result will be similar.
The Virtualized Data Center
This diagram represents the virtual infrastructure discussed above. It shows three VMware vSphere servers with shared storage supporting 12 virtual machines. We explained that the hardware and software keeping the virtual environment costs approximately $10,000 less than 12 physical servers handling the same workload.
Most companies would be interested in reducing CAPEX expenditures when building their infrastructure. But, we also understand that most companies have a heavy investment in physical hardware. Therefore, it becomes a little harder to move to virtualization unless a significant hardware refresh or a new project justifies multiple purchases. So, let’s look at some of the other reasons to consider implementing a virtual infrastructure.
First of all, we need to mention the additional savings in the data center directly related to virtualization. Depending on who you listen to (and where your data center is located), it costs between $400 and $1,800 per year to operate a physical server in the data center (considering HVAC, power, rent, etc.). In the example above, this Bank would save between $3,200 and $14,400 per year in operating costs. Over three years, this would amount to a significant (OPEX) savings.
Second, virtualized servers (VMs) are more accessible to backup and recovery than physical machines. Because the Hypervisor separates the hardware from the VM, once separated, the VM can be moved to or “recovered” on dissimilar hardware in significantly less time, assuming the new hardware is running the Hypervisor. Recovery time can be reduced dramatically in a virtual environment. And, virtualized servers can be replicated to local or remote storage locations using special software applications like vRanger or Veeam Backup and Replication.
Third, virtual clusters can offer high availability servers, which help improve business continuity. In the Virtualized Data Center example above, VMs can be migrated from one vSphere host to another without downtime using a feature called v-Motion. If one of the vSphere hosts needs maintenance, the four VMs can be moved to the other two hosts while the administrator performs the hardware maintenance. Once the problem is solved, the host can be brought back up, and the VMs can be restored. Because of the existence of Shared Storage, this is possible in the example above.
Finally, virtual technology offers increased operational flexibility. Servers can be managed and monitored remotely with essential management tools. Servers can be provisioned more quickly with VMware technology. And VMware includes other native devices like snapshot and rollback that allow administrators to protect server resources when things do not go as planned.