With all kinds of data centers from smaller on-site server farms to colocation facilities, to distributed services, to massive dedicated data centers, and combinations. There is still a divide between operations and financial departments.

The CFO should defniitely care but make sure to be in alignment with IT and the Business Units around balancing efficiency with availability.

One of the major early drivers of colocation has been when an inhouse data center is unable to deliver additional capacity to the business in a timely manner.

Here are a few questions that could be asked:

* How do different applications, SLAs, application workload and performance affect power costs and usage?

* How much energy could be saved when application demand is low?

* How much excess capacity are we paying for but not actually using?

* How much of the power draw goes toward real, revenue-generating work?

* What does one server really cost to run?

* What kind of operation costs are associated with each application?

Power usage effectiveness (PUE), introduced by The Green Grid in 2007, has emerged as one of the most widespread energy metrics. This an equation of total facility power consumption divided by the power used for actual computing.

 However, there are a few additional factors that should be considered:

* Carbon emissions per server (user, application, transaction, URL or kWh)

* Changes over time (throughout a day, a month or a year) - electricity rates, demand

* Energy consumption and cost per application (server, transaction completed or URL served, user)

* Percent of energy consumption and costs that goes to idle servers

* Revenue per kWh or MWh, or the revenue versus costs for all of the above

Power Usage, the Great Divide and Why the CFO Should Care from The Data Center Journal