So what comes to mind when you think about the performance of your data center environment? I suspect the answer may depend on your area of responsibility. If you’re a facility engineer, maybe your primary concern is the reliability and availability of your electrical and mechanical infrastructure. If you’re a technology operations manager, meeting SLAs for network and data processing availability might be a primary focus.

Those with financial responsibility are probably under enormous pressure to manage ever-increasing capital and operational budgets. And if you’re sitting in the C-suite, you’re likely getting hit from all sides with the added responsibility of managing the image and reputation of your firm. Whatever your area of focus, one thing is clear; actions you take, or don’t take, to address efficiency, innovation, the environment, and a myriad of other data center centric performance issues can have long-term impact on your firm’s reputation and ultimately, your bottom line.

Too Much Talk, Too Little Action?

The topic of performance management has taken center stage at most industry events. Seemingly innumerable technical publications, blogs, and webinars highlight the endless opportunities for improving performance. Yet, despite all the talk, industry progress in driving performance is still far too slow.

Yes, there have been some notable successes. Google, Yahoo!, and Facebook are among the large operators that have recently completed data centers projects that set new performance levels for energy efficiency. Others have made strides in virtualization and cloud computing, driving compute utilization higher and energy consumption per unit of compute lower.

Still, it appears that the industry as a whole has not made a lot of progress. Back in 2009, the EPA’s Energy Star program reported an average Power Usage Effectiveness (PUE) of 1.91 across a portfolio of participating organizations. This past May, the Uptime Institute reported an average PUE of 1.8 across facilities they surveyed. While the trend is moving in the right direction, this performance certainly does not reflect the potential opportunities that exist today.

So why is progress so slow? Is it a lack of opportunity? Unlikely. In fact, many experts agree that most existing data centers could dramatically improve the energy performance of their mechanical and electrical infrastructure by applying relatively straightforward efficiency upgrades or even by just adjusting operational practices.

Far too many organizations continue to operate their IT hardware at unnecessarily low utilization rates or even worse leave equipment energized that is not doing any work at all. Adopting strategies that address these inefficiencies not only reduces energy expense but often can free up stranded capacity allowing the deferment of expensive capacity expansion projects. Sitting on the sidelines waiting for the big new data center project to materialize just doesn’t cut it anymore. Not only will executing smart enhancements to your existing facilities yield immediate benefit, the successes that you achieve will help motivate others in your organization to get involved.

Some may say that lack of funding is responsible for the slow progress. I suspect this excuse is overused. In my experience, when a solid business case is made that demonstrates a real-world performance improvement with a definable payback, management usually jumps at the opportunity to act. All too often, proposals for energy initiatives are filled with unsupported claims or based on sketchy in-house data that have not been objectively validated. This can make it very difficult to get the necessary funding, even when the proposed effort has merit.

Working in the Dark?

While some may have a good reason to postpone acting, I suspect that in most instances decision-making based on perceptions rather than facts inhibits real progress. Whether it’s “we don’t have the dollars,” “our data center is too old.” “our operation is too complex,” or any of a thousand other excuses, inaction is often justified by a lack of accurate performance data coupled with poor communication across the many silos of stakeholders. Operators that lack the detailed data necessary to empirically validate performance often find themselves creating lots of general “rules” to guide their operations. As long as everyone keeps to the rules there is little need for much discussion. While it is crucial to have solid processes to manage your operation, it is important to not allow those processes to inhibit challenges to the status quo. Arming people with accurate data and encouraging an environment of collaboration across stakeholder silos will go a long way to ending the communication gap so common in our industry.

fig 1
Figure 1

If you think the communications gap across the various subject matter experts within your firm is wide, consider that the various industries do not treat similar challenges in the same way because they face different market pressures. In the financial services data center arena, a sudden spike in the financial markets can test the boundaries of processing capacity and even a small hiccup can result in major financial losses, regulatory penalties, or worse, can scare clients right over to your competitors. Protecting against these exposures will always be at the top of the priority list, even if it is at the expense of efficiency. Other industry silos have addressed this type of challenge by instantiating capacity on a public or private cloud to pick up the slack on an as-needed basis. A highly regulated environment can make decisions about the cloud very challenging as operators must be sure that access to highly sensitive data is secure, both when it is in use and after that capacity is released for other uses.

If you’re an internet or social media service provider, you have a different set of challenges. Yes you need reliability, but often your biggest challenge is keeping up with the capacity required to support an ever-growing user base. Those massive and continuously growing compute volumes, along with their associated data center operational and capital expenses, are often the single largest contributor to your bottom line. While you may not have the same client challenges or regulatory limitations of a financial organization, your massive scale often puts you in the cross hairs of those driving environmental policy.

The list goes on. The government, academic institutions, research, manufacturing, and the general commercial sector all have unique business and operational requirements. All too often, these differences are used to explain why what the “other guys” are doing won’t work for their operations. The flood of questionable performance data disseminated by sales and marketing departments has added to the confusion. Metrics such as PUE that, when properly applied, can be very helpful in describing high level energy performance, are often unfairly disparaged. Yes, people misuse them. Yes, PUE doesn’t address overall compute efficiency. But when properly applied, PUE can set the stage for a deeper dive into the details by providing a baseline along with a frame of reference to measure progress against.

Building a Business Case

Whether your goal is to start-up a new data center performance management process or reinvigorating a stalled one, it’s important to take a step back and reevaluate what you are trying to accomplish. While implementing one-off energy initiatives certainly can be beneficial, the most successful energy and environmental programs are the ones that are driven by a business plan that clearly articulates a value proposition in alignment with the strategic goals of your firm.

Creating a successful business plan will require some effort. A common mistake is to wait too long to get started and allowing year-end capital planning deadlines to drive the process. This frequently results in rejection of the project due to insufficient backup. Another misstep is to assume the acceptance of other stakeholders working in the environment. No one likes surprises, particularly if your initiative could potentially impact their world, or worse, steal the thunder from their efficiency initiatives. Engaging other stakeholder groups early in the process can be pivotal to gaining approval for a project. Finally, if your area of expertise is more technical then financial, try and find an ally with the business acumen to help you nail the commercial aspects of your plan.

A solid business plan will accomplish several things. First and foremost, it will demonstrate that the proper level of homework was done thereby giving credibility to the process. Second, since it is rare that a single person or department will have exclusive control over the data center, it will provide the information necessary for all stakeholders to evaluate project value and impact from their area of focus. Lastly, since the format engages the various stakeholders early in the process, it will have a far greater chance of gaining grass roots support, thereby making it much easier to get the approvals necessary to proceed.