Data Center ArchitectureVirtualization

4 Effects that Server Virtualization Can Have on Data Center Infrastructure

It’s well understood that virtualization technology has enabled IT groups to take great strides in increasing the utilization rate, scalability and flexibility of their servers. What’s more, it is the technology engine behind cloud computing.

What’s less well understood are the effects virtualization has on data center physical infrastructure (DCPI). To help clear the air, Schneider Electric’s Energy University is now offering a free course, “Virtualization: The Engine Behind Cloud Computing.”

The course looks at four effects that virtualization has on DCPI, namely:

  1. Creation of more high-density areas and hot spots
  2. Potentially detrimental effect on power usage effectiveness (PUE)
  3. Dynamic IT load swings
  4. Lower redundancy requirements

1. The rise of high density

Virtualization enables organizations to achieve server consolidation ratios of 10:1, 20:1 or even higher. The whole idea is to achieve higher CPU utilization rates on the remaining physical hosts. As a result, those hosts draw more power. What’s more, they tend to get grouped together in ways that create localized high-density areas and, hence, hot spots.

The course goes through some of the options for dealing with this scenario, including spreading out the high-density equipment throughout the data center or consolidating high-density systems into a pod with dedicated cooling and/or an air containment system.

2. Reduced IT load effecting PUE

Even though high-density servers may draw more power than those they replaced, there are far fewer of them after a consolidation effort. As a result, server virtualization efforts typically result in a much lower electric bill.

It may seem non-intuitive, then, that the data center’s PUE rating can actually get worse after a virtualization and server consolidation project. The reason, as the course explains, is what’s known as “fixed losses,” which is the power consumed by power and cooling systems regardless of what the IT load is. As the IT load shrinks, these fixed losses will become a greater percentage of the overall data center energy use – which means PUE will worsen.

The course goes through some options for optimizing the power and cooling infrastructure after a server consolidation, to get it the aligned with the new, lower IT load.

3. Dynamic IT loads

One of the benefits of virtualization technology is the ability to shift loads dynamically from one server to another as demand or other factors, such as a server failure, dictate. But this sudden – and increasingly automated – creation and movement of virtual machines requires careful management and policies that take into account the status and capacity of DCPI down to an individual rack level. Failure to do so could undermine the software fault-tolerance that virtualization brings to cloud computing.

The Energy University course explains how data center infrastructure management (DCIM) software can monitor and report on the health and capacity status of the power and cooling systems, and keep track of all the relationships between the IT gear and the physical infrastructure to ensure IT loads aren’t moved to areas that can’t handle them.

4. Lowered redundancy requirements

Finally, the course uncovers one of the lesser-known benefits of server virtualization: it can lead to a reduced need for redundancy in the physical infrastructure. With the ability to dynamically shift loads, a well-managed virtual server implemention brings with it a high degree of fault-tolerance for both the servers and the applications that run on them. Workloads, entire virtual machines, and virtualized storage resources can be automatically and instantly relocated to safe areas of the network when problems arise.

This level of fault tolerance may reduce the need for a highly redundant (i.e., 2N or 2N+1) power and cooling system in a highly virtualized data center. If, for example, the failure of a particular UPS does not result in business disruption, it may not be necessary to have a backup, redundant UPS system for the one that failed. The course explains how companies can save money in various scenarios, depending on their requirements and server configuration.

Probably your IT group has already embarked on a server virtualization project of some sort. Make sure your DCPI is in lock step by taking the free course, “Virtualization: The Engine Behind Cloud Computing.”  You’ll find it in the College of Data Centers at Energy University.


No Responses

Leave a Reply

  • (will not be published)