IT departments today are all about speed – adding more applications, faster, to satisfy voracious global appetites for mobile, social media, streaming video and other resource-intensive applications. These applications must be available to satisfy employees, customers and stakeholders, regardless of their impact on critical infrastructure performance.

If growth is stretching your company’s IT resources too thin, you must consider what needs to be done to ensure availability. IT departments are facing continued pressure to lower costs, and so taking proactive steps will require you to build a solid business case when deciding how to expand.

A variety of solutions are available for upgrading and expanding your critical infrastructure - including building a new data center in some cases. But how do you know which option is best? Whatever path you take, consider the following questions when building your business case.

Have you assessed and optimized existing critical infrastructure?
Most data centers offer much room for improvement, and these questions will help you discover if yours is one of them:

• * Are you following efficiency best practices that can add capacity?
• * Do you know that you’re getting everything you should from your assets; what are your utilization rates?
• * Do you have dead servers sitting around wasting capacity?
• * What percentage of servers have you virtualized; can you boost that?
• * Are your racks as dense as they could be?
• * Are you/can you take advantage of new, more efficient technologies?
  
  If you’re behind on adopting best practices, such as using power management features on servers and economization for cooling, you could potentially be throwing away the chance to significantly increase efficiency, which gives you more power and cooling capacity.
  
  Server virtualisation is an effective way to gain capacity in an existing data center. If you’ve already virtualized, there’s every chance you may be able to go further. For example, increasing from 30% server virtualization to 60% can provide a 29% reduction in data center energy consumption. This will provide you with additional power and rack space for growth.
  
  Data center infrastructure management (DCIM) can help you increase virtualization rates by providing visibility into how virtual servers are deployed and into the infrastructure capacity supporting those virtual servers. Therefore, infrastructure capacity can be fully realised without risking over provisioning.
  
  Companies have saved as much as 50% of data center energy just by virtualizing and turning off what they aren’t using. Getting a handle on this wasted energy saves money and capacity. Do you know if all your servers are serving a purpose?
  
  Increasing rack density is an additional opportunity to increase efficiency and free up capacity. Many businesses have increased density to 6-8kW per rack, but few have pushed it to 12-14kW per rack. More servers per rack means fewer racks taking up floor space - or more racks to take on new servers. On the cooling side, if you upgrade your cooling systems to newer, more efficient technologies you can get 50 to 100% more cooling capacity.
  
  In short, a comprehensive assessment of your existing infrastructure will find stranded and/or wasted capacity and inefficiencies that can be addressed by retrofitting.
  
  Colo or cloud infrastructure?
  If optimizing your critical infrastructure doesn’t recover enough capacity, consider using someone else’s infrastructure:

• * Using a colocation facility for overflow; some newer colos offering dedicated suites provide more control than co-los with cages of shared resources.
• * Moving to the Cloud; one strategy is to keep critical applications in-house, put the rest in the cloud, taking a balanced approach to get the best compute and space deal you can.
• 
  When should you consider building new?
  With all the opportunities available, new builds are a last resort for many. Building new is expensive and difficult. A new data center can cost in the range of $17M per megawatt, though there are ways to push this down to $12M per megawatt (for example, through UPS utilization improvements).
  
  Factors such as getting a loan, finding a location and determining if you can get fiber optic cable all come into play. It can take as long as three years before the new space is operational, and managing logistics for a move can be challenging.
  
  Still, there are times when building new is a logical forward step:

• * Your business growth is through acquisition, and you need to consolidate a number of acquired data centres. You may have to build new to accommodate them.
• * You don’t have a disaster recovery site. You can build new and use the old facility for this purpose.
• * Your equipment is so old and poorly maintained that it’s not worth merely upgrading. Building new is an opportunity to take advantage of advanced, more efficient technologies.
  
  Whatever route you decide to take, it’s certainly worth giving all the options full consideration in order to ensure that you are maximising cost and energy efficiency, and planning carefully to map your infrastructure changes against business needs.