Manage Learn to apply best practices and optimize your operations.

Think before you build with data center predictive modeling

If you're thinking about a new data center for your business, data center predictive modeling can help you cover all your bases right up front.

There is more involved in setting up a data center than bringing a bunch of computers into an empty room and running some cables. A data center's job is to provide critical resources to the computing devices that will be housed there. Data center modeling can help you decide what to buy.

At its simplest, data center modeling can be thought of as data center resource management. A data center model typically maps equipment placement within the data center, but also accounts for the use of things such as the following:

  • Server rack space
  • Floor space
  • Electricity (and electrical outlets)
  • Heat dissipation
  • Equipment weight distribution
  • Network connectivity

It is easy to see how a data center model could be useful when setting up a brand new data center, but the modeling process is most effective when used on an ongoing basis. The modeling process can be performed manually, but there are also a number of data center modeling applications that are available from various vendors. Regardless of whether you model your data center manually -- using things like Visio or AutoCAD drawings -- or you purchase a dedicated data center modeling application such as CenterOS or Rackwise, there are several data center modeling tasks that should be performed regularly.

Capacity management

The most obvious use for ongoing data center modeling is that your model can help you with capacity management. For example, if someone in your organization needs to purchase a new server, then your data center model should be able to tell you if you have rack space available and if there are power and network cables in close proximity to the rack. After all, it would be embarrassing to install a new server only to find out that you can't plug it in without tripping a circuit breaker.

Asset tracking

Another use for data center predictive modeling is asset tracking. A data center is typically one of the largest investments a company makes, so it is important to be able to keep track of the hardware that is used within the data center. Whether you perform data center modeling manually or by using specialized software, you should record the serial number and location of each piece of hardware.

Of course, tracking serial numbers and physical locations alone is of limited usefulness. This type of asset tracking can help you to confirm a piece of equipment's physical whereabouts, but asset tracking often involves keeping track of much more detailed information.

I recommend that for each piece of equipment in the data center you record the following pieces of information:

  • Manufacturer
  • Make and model
  • Serial number
  • Physical location
  • Configuration details (such as a server's memory, CPU and hard disk configuration)
  • The purchase date
  • The original purchase price
  • The name and contact information of the vendor from which the equipment was purchased
  • Warranty information
  • The vendor who holds the maintenance contract (and that vendor's contact information)
  • The maintenance contract's expiration date

If possible, it is also a good idea to keep a record of any parts that have been upgraded or replaced, as well as the cost of those parts.

Keeping these types of detailed records for each component within the data center can greatly simplify the process of getting the cost of a repair covered under a warranty or maintenance contract. More importantly, you will need this type of detailed information if you ever have to file an insurance claim for damaged or stolen equipment.

Hardware lifecycle management and planning

Perhaps the most important ongoing use for data center modeling is hardware lifecycle management. Computer hardware has a limited lifespan. Although server virtualization technologies have helped with this to some extent, it is inevitable that data center managers will have to replace old equipment with newer equipment. Likewise, evolving business needs can also result in new hardware acquisitions.

More on capacity planning

Big data requires better capacity planning

Scale up or scale out? What works for you?

Capacity management for virtual resources

Whether an organization is investing in new hardware or replacing aging hardware, the implementation process has an impact on the data center. Every new computer consumes electricity, switch ports and floor or rack space. Likewise, every computer gives off heat as it operates. As such, there are two things that your data center model should do for you when it comes to adding or replacing computer hardware.

First, your data center model should help you to spot deployment issues before the new hardware is purchased. For example, your model may indicate that you need to purchase additional rack space or beef up your cooling system before you can deploy the new hardware.

Second, your data center model should allow you to experiment with various what-if scenarios. For example, a good data center model should be able to help you to answer questions such as the following:

  • If I move equipment to free up floor space, will I have enough electricity in that part of the data center?
  • How much power will I save if I consolidate five physical servers into a single virtualization host?
  • What impact will a new disk storage appliance have on the data center's temperature?

A good data center model should be a data center manager's go-to resource for any changes that occur within the data center. Regardless of how the model is created, it should be able to track existing assets and help plan for new ones.

ABOUT THE AUTHOR: Brien Posey is a seven-time Microsoft MVP with two decades of IT experience. During that time, Posey published thousands of articles and wrote or contributed to dozens of IT books. Prior to becoming a freelance writer, Posey served as chief information officer for a national chain of hospitals and health care facilities. He also worked as a network administrator for some of the 'nation's largest insurance companies and for the Department of Defense at Fort Knox, Ky.

Dig Deeper on Data center capacity planning

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.