In this day and age, everyone wants to become greener. Motivations vary: reduction of operating costs by reducing energy consumption, reduction of carbon footprint and the impact of pending carbon taxes on high-energy users, leave the planet in good shape for our kids, PR, whatever.

Square foot – square foot, data centers consume more energy than about any other building on the planet. In the past, availability trumped all issues for data center operations teams and rightfully so. Now we’re questioning if we can maintain our system uptime while radically decreasing the energy those systems consume.

In the past five years, we’ve seen the abandonment of availability as the sole bludgeon and über metric of senior IT operations management. Face it, if your data center doesn’t exceed the availability needs of your enterprise, you’re riding the pine.

Data centers have grown up as slaves to system availability and uptime. That’s been a good thing, and we’ve seen the “factory” evolve through various steps that have hardened the hardware, network, building, utility infrastructure as well as the applications. The delivery model in some markets celebrates the normalization of data center infrastructure into a homogeneous deliverable and then monetizes that product as a consistent client backplane and expectation in the real estate field. Super, so once you’re over that hurdle, where do you go?

The fact is that well-engineered systems and buildings have always respected their life-cycle costs. My contention is that the data center is no more complicated than an industrial factory that makes information. And any well-run factory is optimized for efficiency and output, when it works as an integrated system. While improvements will undoubtedly continue for availability improvements, they are becoming increasingly granular in their ability to meaningfully move the mathematics of measured application availability. The attention and capital of the industry will naturally move elsewhere. And that elsewhere is efficiency.

For energy, this metric can be rendered as DCiE or PUE. What we see beginning to happen is a conscious decision on the types of utility systems deployed and what they cost to operate. This will take the workload of the IT platforms and couple it to both a very accurate monitoring system – a more real-time active power management of the platforms at the receptacle level in the data center

As this blog evolves, we’ll discuss the varying area of efficiency. For the next few blogs, we’re going to focus on energy and this application of system selection, critical power system efficacy, and monitoring of the IT loads themselves?

The thoughts we’ll be dealing with in the coming weeks are:

• What are the true data center efficiency figures for the typical electrical and mechanical systems employed? And is there a “sweet spot”?
• Are given systems or design better than others?
• How do we effectively measure the energy consumed by the IT systems? What are some of the approaches and pro and cons?
• What types of electrical controls and distribution systems are available to manage my utility costs?
• Can the monitoring solutions at the rack or system level provide me failure prediction for the platform they are monitoring?