10-Time Microsoft MVP Brien Posey Discusses Hardware Considerations for Private Clouds

Hardware Considerations for Private Clouds

— Brought to you by 2X Cloud Computing guest blogger Brien M. Posey —

When it comes to developing a private cloud, you will likely have to put a great deal of planning into the hardware that makes up your cloud infrastructure. Your hardware must provide three things:

Although each of these considerations carries its own implications, they are also complementary to one another. This has a lot to do with the fundamental nature of cloud services and with the way that the underlying hardware and software work together to deliver private cloud functionality.

Let’s start by talking about performance. In years past, server hardware was matched to the application (or application set) that was going to be running on the hardware. It was no coincidence that the hardware was referred to as a server, but the application software was also sometimes referred to as a server.

In a private cloud environment however, performance must be thought of differently. Private clouds are based on server virtualization. In a virtual datacenter, applications are not matched to server hardware. Instead, physical hardware is assigned workloads that are made up of multiple virtual machines. As such, ensuring performance means making sure that the workload that is assigned to the hardware does not exceed the hardware’s capabilities.

Another requirement for any private cloud is reliability. The private cloud must be constructed in a way that ensures that the failure of any component will not result in an outage. Obviously, redundancy plays a major role in establishing reliability and there are many different techniques for implementing redundant hardware. One of the main techniques that is used however, is host server clustering.

As you are no doubt aware, a cluster of host servers allows workloads to fail over to another node within the cluster in the event of a host server failure. While this failover capability is usually implemented for the sake of reliability, the ability to dynamically shift workloads from one host to another can also help with performance. Automation software can be put into place that will dynamically move virtual machines from one host to another if a particular host begins to run low on hardware resources.

The third concept that must be addressed when planning private cloud hardware is that of scalability. As time goes on, your private cloud will likely be required to run an ever increasing number of virtualized workloads. That being the case, the private cloud must be built in a way that will allow it to grow as the demand for private cloud resources increases.

The concept of scalability ties in directly with performance and reliability planning. Obviously higher end hardware can handle a greater number of virtualized workloads, but there is more to it than that. Host server clusters can only accommodate a limited number of cluster nodes. The exact number of nodes that can be used is determined by the underlying operating system. For example, failover clusters that are built on Windows Server 2012 can accommodate up to sixty-three nodes. Other operating systems tend to support far fewer cluster nodes however. For example, Windows Server 2008 R2 has a maximum cluster size of sixteen nodes.

Because a failover cluster can only support a limited number of nodes, it is often better from a scalability perspective to initially create your cluster using a relatively small number of high-performance nodes than to use a large number of low performance nodes. Of course even the best planned clusters may eventually become inadequate to service the anticipated demand. When this happens, the solution is typically to distribute workloads among a series of independent clusters that collectively work together to provide cloud resources.


Performance, reliability, and scalability must all be taken into account when planning the hardware that will be used in your private cloud. Although each of these factors should be assessed independently, they all have an impact on one another and should therefore also be considered collectively.

About Brien M. Posey

Brien Posey is a ten time Microsoft MVP with two decades of IT experience. Prior to becoming a freelance technical writer, Brien served as CIO for a national chain of hospitals and healthcare facilities. He has also worked as a network administrator for some of the nation’s largest insurance companies and for the Department of Defense at Fort Knox.

Since going freelance in 2001, Brien has become a prolific technical author. He has published many thousands of articles and numerous books on a wide variety of topics (primarily focusing on enterprise networking). In addition to his writing, Brien has provided consulting services to clients and speaks at IT events all over the world.

About 2X Software

2X Software is a global leader in virtual desktop and application delivery, remote access and cloud computing solutions. Thousands of enterprises worldwide trust in the reliability and scalability of 2X products. 2X offers a range of solutions to make every company’s shift to cloud computing simple and affordable.

For additional information, visit www.2x.com or contact Charlie Williams by email cw@2x.com, phone +356 2258 3800.

Leave a Reply