Jakub Jirsk - Fotolia
Take control of Hyper-V power management for cost efficiency
Organizations can better manage Hyper-V power consumption through various means, such as using Hyper-V power plans, server cores and SCVMM capabilities.
Power is one of the largest ongoing costs associated with operating a data center. The average annual cost per kilowatt of required power can be as high as $26,495 in small data centers, according to Data Center Knowledge. Given the high cost of power, it is clearly in an organization's best interest to do what it can to reduce data center power consumption. One of the best approaches is for organizations to take control of their Hyper-V power management and focus on reducing the power consumed by Hyper-V VMs.
Larger data centers such as those over 50,000 square feet of space typically enjoy far lower power rates. However, they still spend an average of $5,467 per year for each kilowatt of required power. By readjusting their Hyper-V power management strategy, organizations can not only lower costs, but also improve system performance.
Explore the different Hyper-V power management plans
One of the first things that organizations should look at is the power plans being used by their VMs. The Windows Server OS supports a number of different power plans, providing organizations with the flexibility to choose the plan that best suits their needs.
Balanced power plan. Windows Server uses the Balanced power plan by default and it is adequate for most use cases. This plan balances a virtual server's performance with its energy consumption. It does this by throttling CPU to match workload demand.
High Performance power plan. Windows Server also supports a High Performance plan. This plan locks CPUs into their highest supported performance state and unparks all CPU cores. Though this mode might be suitable for an organization's highest demand workloads, it also consumes more energy than the Balanced plan.
Power Saver plan. Another option is to use the Power Saver plan, which prevents CPUs from being used beyond a percentage of their maximum capability. In addition, the Power Saver plan enables other energy-saving features, such as core parking. This plan is best suited for low-performance workloads or for organizations that must greatly reduce their power consumption and thermal output.
It is difficult to quantify how these plans impact the actual operating costs for organizations given the fact that savings vary dramatically based on the underlying hardware and workload.
Additional strategies and techniques boost cost savings
Although choosing the most appropriate power plan is a good first step to taking control of Hyper-V power management, there are other things that organizations can do to optimize their VM power consumption. One commonly used strategy is to customize the default power plans. For example, organizations might configure a power plan to turn off inactive hard disks or the display when it is inactive. These small changes can reduce power consumption, even on VMs.
Another way that organizations can reduce power consumption is to perform server core deployments. Server core deployments lack the Windows Desktop Experience. Because VMs do not have to perform the additional processing necessary to support a GUI interface, they use less power. Performing server core deployments can also help to increase VM density, thereby reducing the total number of physical servers required to run organizations' collections of VMs.
Similarly, assigning a single server role to each VM can help keep a VM's power consumption in check by reducing both the processing requirements and the overhead associated with patch management.
Finally, if organizations use System Center Virtual Machine Manager to manage their Hyper-V VMs, then they should be using SCVMM's Power Optimization feature. The Power Optimization feature utilizes SCVMM's Dynamic Optimization capabilities, which seek to balance workloads by live migrating VMs to underutilized servers. The Power Optimization feature takes this a step further by automatically condensing VMs onto fewer hosts during periods of light use, such as after-business hours, and then shutting down unused Hyper-V hosts. When heavier activity resumes, the unused hosts are automatically powered on and some of the VMs are then live migrated back to those hosts.