While data center energy efficiency strategies have reached high rates of adoption, many organizations and IT professionals fail to realize that improving energy consumption is only one aspect of managing an efficient data center. The practice of data center performance optimization maintains that an optimized data center actually achieves three types of efficiency — energy consumption, resource utilization and process efficiency — enabling organizations to achieve an optimal balance not only of energy consumption, but of infrastructure resource utilization and process efficiency as well. Fueled by media reports and research from organizations such as the Green Grid, data center energy efficiency initiatives have quickly achieved extensive adoption. Yet while energy efficiency has received widespread attention for its ability to reduce costs in the data center and support green initiatives, it should not be the only priority of IT management. As IT management is increasingly focused on not only cost savings, but doing more with less, many are realizing that optimizing resource utilization presents a powerful opportunity to address both of these concerns. Effective management of resources enables organizations to proactively manage capacities and ensure that critical power and cooling resources are sized correctly and balanced with computing demand, without jeopardizing availability.
Better resource utilization can also help an organization navigate difficult economic times by delaying capital investments, through improved utilization of all of the available resources. Improved management strategies cannot only enhance data center resource utilization at the data center infrastructure layer, but also at the IT infrastructure layer, which is comprised of compute, storage and network equipment—and the size of the opportunity is similarly large.
A typical example is the presence of unproductive, or “ghost” servers in a data center. Inactive servers can consume 65 to 75 percent of the power of a working server, representing not only wasteful energy consumption, but resource inefficiency, as the servers could instead be re-commissioned to perform valuable work. Improved visibility and management at the IT infrastructure level can help organizations identify, remove and reallocate servers to tasks that contribute more efficiently to the IT operation. Additionally, effective management of data center resources not only reduces costs by improving resource allocation, but also mitigates risks by providing enhanced visibility into data center resources.
Another troubling trend is that many organizations do not know the operating power density of their data center. Failure to appropriately manage resources by accurately tracking power consumption (densities) and capacities can lead to the over-provisioning of data center resources to hedge risk, or inadequate provisioning due to lack of insight into demand. In either case, money is wasted by paying for resources that are not needed (not fully utilizing infrastructure resources), or by threatening costly outages.
Process Efficiency Lags Farthest Behind in Adoption
Resource utilization is not the only overlooked target for cost reductions in the data center. Amidst reduced staff levels and ever-increasing demand on the data center, process efficiency is also critical to optimizing performance as energy and resource efficiencies. Among the three efficiencies, process efficiency lags farthest behind in adoption, particularly in configuration management (data center documentation).
The task of managing configuration information is a top opportunity to improve staff effectiveness. Some data center managers are using three to five systems to document their infrastructure; while others are updating documentation in the data center manually. These issues can only lead to the conclusion that data center personnel are not being used as efficiently as they could be, effectively costing the organization money.
Using disparate systems that are not updated within a well defined process is also an inefficient use of time and people resources when solutions for managing the data center from a single system are available. Furthermore, the inaccuracies of manual data input can also threaten service levels. Many data center managers experience data accuracy problems related to human error when gathering data, meaning that not only is staff effectiveness reduced under existing data collection strategies, but availability is also threatened. This is a critical, often overlooked, hidden cost to the organization. Inaccurate documentation (configuration information) leads to bad decisions, costly rework, increased mean-time-to-repair and potential outages.
Conclusions
The practice of optimizing data center performance to manage the balance of service levels and costs can be viewed as the management of three types of efficiency – energy consumption, resource utilization and process efficiency. Organizations are failing to appropriately manage all three efficiencies, effectively operating in an imbalanced state, and missing opportunities to achieve efficiency on a larger scale. Furthermore, they are preventing the efficiency initiatives they do have in place from achieving their full potential.
As economic conditions continue to compel organizations to not only reduce spending but to achieve maximum utilization of all resources available to the business, all three pillars of efficiency – energy, resource and process -- will become increasingly important to effective management of the data center.
About the author
Steven Yellen brings more than 20 years of extensive marketing and technology experience to his role as Vice President of Marketing and Strategy at Aperture Technologies, an Emerson Network Power brand. He has authored numerous white papers and technical articles, and has frequently spoken at industry conferences and seminars on both Data Center Management and Sales and Marketing Integration.
