DCOI SUMMARY
The US Office of Management and Budget dictates that federal data centers must meet a PUE rating of 1.5 or better and must install data center infrastructure management (DCIM) software by 2018. The corresponding initiative, known as the Data Center Optimization Initiative (DCOI), encourages existing data centers to immediately transition to the usage of DCIM software and to begin incorporating "green" IT technology to meet the 2018 deadline for identified data center optimization targets. Further, new federal datacenter builds are required to meet a more stringent PUE rating of 1.4 during inception. This mandate stresses the need for both competent DCIM software and energy practices to achieve either of these PUE objectives, and it views DCIM software not as an optional extra but as a vital component of the data center to meet such goals.
What is PUE?
PUE (Power Usage Effectiveness) is a statistic that measures the efficiency of energy usage within a data center. Developed by the Green Grid organization in 2006, PUE is currently the most widely used measurement to determine the effectiveness of power usage within the datacenter environment and is used as the de facto standard in benchmarking any data center against ideal efficiency targets and required goals. Companies such as Google strive to meet or exceed low PUE targets such as 1.21 or below and actively promote better energy practices to drive that goal even lower over the long-term.
HOW IS IT CALCULATED?
PUE is calculated by dividing the total amount of energy consumed within the data center space (the total Facility Energy) by the total amount of energy consumed by computing equipment (such as servers, network devices, storage arrays etc.) The energy consumed by the data center would include all powered equipment that are not considered computing devices, such as cooling, lighting and power distribution equipment. This calculation should always be 1 or higher, with a perfect calculated value of 1. For example, a data center that consumes 1,000 kilowatts of power but whose computing equipment consumes 800 kilowatts would yield a PUE rating of 1.25 (1,000kW / 800kW). A typical data center PUE rating would be considered efficient at a value of 1.5 or lower; yet, many companies seek to attain ratings far closer to 1.2 or 1.1.
HOW CAN NORLINX ASSESS A DATA CENTER'S PUE?
NORLINX Global Site Management(GSM) collects power consumption data from a variety of power equipment (or measurement points) in the data center through the usage of its properitory "poller" services running on one or multiple platforms. The intervals of data collection are customizable by authorized users, and the power data can be evaluated based on user-specified validation rules. Thus, GSM can collect data from such power equipment, from the outlet level to the power feeds into a cabinet, and then aggregate power load data for all the computing devices inside the cabinet(s) including the supplied power feeding the cabinet(s) as well. From this data, GSM can provide an accurate PUE calculation in real-time, based on live data from the devices supplying infrastructure power to those IT devices consuming power within the cabinet(s). Through GSM's power aggregation capabilities, PUE can be calculated from the smallest measurable elements within the data center to the data center itself as a whole.
HOW CAN NORLINX HELP FEDERAL DATA CENTERS ACHIEVE THE REQUIRED PUE RATINGS?
Much of the groundwork to achieving a target PUE goal of 1.5 or better is accomplished by simply exposing the current power environment to Facilities and IT personnel. Those responsible for the energy efficiency of a data center must first know exactly what power is being consumed in the data center (from both the Facilities and IT sides) and then where any anomalies lie. GSM is designed to provide both visual indicators and customizable alerting mechanisms to communicate this information. GSM can isolate exactly which devices are in danger of exceeding efficiency limits and which devices are no longer serving a valuable function (and yet consuming power). For example, a user can compare various models and configurations of IT devices within GSM to identify those models which consume the most power or that run most inefficiently. Armed with that knowledge, that user can then remedy that situation by either consolidating such devices into lower-consumption blade systems or by purchasing "greener" variations of that equipment. Time-specific power consumption issues can just as easily be identified through a variety of historical trend reports within GSM. GSM can also take advantage of a host of smart sensor technologies available on the leading monitored power devices to fine-tune utilization and control of power efficiency and consumption down to the outlet or connected device. GSM also provides tools to optimize the balancing of power loads within the data center, down to the breaker or outlet. The more balanced the power load is across all phases within a cabinet, the more energy efficient the cabinet can become.