Inside the capacity management process: Resource supply and demand

Keep your servers up to snuff by ensuring that resources have been optimally allocated. Understanding resource supply and demand is the key to a strong capacity management process.

This article can also be found in the Premium Editorial Download: Virtual Data Center: How to make the transition from physical to virtual and back:

In a virtual infrastructure, having a strong capacity management process in place is critical to maintaining solid

server resource allocation.

When you begin using virtualization, capacity planning charts the course for resource balancing. Capacity management is the continual process of maintaining optimal virtual machine (VM) capacity.

Many IT professionals also associate the capacity management process with performance management, which makes sense. The goal of performance management is to identify parts of an infrastructure where performance isn’t up to snuff and to tweak resource levels to keep VMs operating at peak levels.

Although the capacity management process uses the same metrics as those used for performance management process, capacity management has a different objective from the goal of performance management. If performance management is about the behavior of today, you might say capacity management is about the behavior of tomorrow.

The economics of the capacity management process
One of the easiest ways to explain capacity management is by invoking a bit of high school (or college) economics. The capacity management process begins with a concept you’re probably familiar with: supply and demand.

In economics, supply and demand refers to how much of a product is available and how much of the product is sought by consumers. The same forces are at work in a virtual infrastructure -- only in this case, the product is physical resources, and the consumers are virtual machines.

Most administrators want to squeeze as many VMs as possible onto as few physical hosts as possible while also ensuring that every VM has an acceptable level of performance . But reaching optimal server consolidation requires a strong capacity management process, and that’s a constant push-and-pull routine: balancing the supply of available physical resources with the demand pressure exerted by running VMs.

What you may not realize is that you can actually gather supply-and-demand metrics from your virtual platform’s management console. Inside the VMware vSphere Client, for example, look at the Virtual Machines tab after highlighting a virtual host. It features a list of VMs with information on their resource utilization.

Interpreting capacity management metrics
Those capacity management metrics help determine resource balance (or imbalance). Note that the numbers are integers, or whole numbers, rather than percentages. Capacity management in a virtual infrastructure can be difficult to grasp because traditionally, most admins have viewed metrics as percentages instead of whole numbers. During the days of physical servers those percentages might have served us well, but whole numbers are more useful for a virtual capacity management process.

So what’s the meaning of these whole numbers? They indicate the supply of host resources, and it takes a simple calculation to determine the demand. Once you know where to find these numbers and how to interpret them, the capacity management process gets easier.

Let’s take processing CPU, for example. In the Virtual Machines tab, a column labeled Host CPU-MHz shows the quantity of Mhz each virtual machine currently demands. For each VM, the amount obviously varies depending on how great a VM’s processing needs are. You can add the MHz demands for each VM on a host to determine how many MHz (or, probably GHz) the total collection of VMs needs. Now you have your demand metrics.

To determine the supply number, check the Summary tab for the host. There, the Resources box has a bar graph marked CPU Usage. Next to the bar graph is a measurement called Capacity, which is the number of CPUs in the host, multiplied by each processor’s GHz contribution -- or the supply of processing resources a host contributes. If a host has four processors that contribute 2.26 GHz each, for instance, your server supplies 9.04 GHz of processing.

You now know the value of a VM’s host supplies and processing demand. By subtracting demand from supply, you can determine how much capacity remains available on a physical host. Voiìa!

This capacity management process is simpler and far less costly than many solutions that calculate the metrics for you. Other factors can affect capacity, including the change rate of these values and host processing reserve, but this simple math provides a quick and accurate capacity measurement.

For a complete capacity management process, repeat the calculation for networking, storage and memory resources.

About the expert
Greg Shields is a partner and principal technologist at Concentrated Technology, an IT analysis and strategic consulting firm. Follow him on Twitter @ConcentratdGreg.

This was first published in January 2011

Dig deeper on Capacity planning for virtualization

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

SearchVMware

SearchWindowsServer

SearchCloudComputing

SearchVirtualDesktop

SearchDataCenter

Close