The data center has become a strategic business asset that continues to grow in importance. Data center issues are now business issues. At a time when the criticality of the data center continues to increase, in many cases, IT budgets are remaining flat or decreasing. The result is infrastructure capacity in data centers is often stranded and misunderstood, while user demand for IT is increasing. Data center managers find themselves under extreme pressure to optimize operational expenses, justify equipment upgrades, delay capital expansions, and explore ways to reduce energy consumption without risking downtime.
The primary method that people use to measure the energy effectiveness of their data centers is power usage effectiveness (PUE). PUE compares the power used for computing capacity and the overall power used by the data center (PUE = Total Facility Power / IT Equipment Power). The inverse of PUE is data center efficiency, or DCE, according to “Green Grid Metrics: Describing Data Center Power Efficiency,” a white paper issued by The Green Grid association. Currently, the industry average PUE is reported to be around 2.0, but a number of new and proven data center infrastructure technologies are available to advance efficiency and reduce PUE to 1.5 or lower without compromising availability. These improvements enable enterprises to save on energy costs while maintaining high levels of availability.
Using monitoring capabilities
Progressive data centers are entering a new stage of maturity marked by a more proactive approach to management to enable increased efficiency, better planning, and higher levels of service. All of this starts with the physical support systems around a data center, especially power systems. Achieving actionable visibility into operations requires the ability to collect, consolidate, and analyze data across the entire data center infrastructure layer. That requires monitoring capabilities.
There are four key actions data center managers can take with integrating monitoring capabilities into power systems that can help improve efficiency and advance PUE (see Figure 1).
1. Monitoring power usage. With power densities and energy costs rising, the ability to monitor energy consumption is essential for advancing PUE. Where power is measured can have an effect on how efficiency is measured.
For a comprehensive picture of data center power consumption, power should be monitored at the UPS, the room power distribution unit (PDU), and within the rack. These data points should be consolidated and reported within an enterprise data center infrastructure management (DCIM) tool that can be shared by both facilities and IT management. DCIM tools (if specified early and fully integrated) can provide an overarching view of the data center, which will provide insights into how effectively or ineffectively the computing power of the data center is being used (see Figure 2). Specifically, measurements taken at the UPS provide a base measurement of data center energy consumption that can be used to calculate PUE and identify energy consumption trends. Monitoring the room PDU prevents overload conditions at the PDU and helps ensure power is distributed evenly across the facility.
The best view of IT power consumption comes from intelligent PDUs inside racks, which feature integrated monitoring and control capabilities to enable continuous power monitoring and feedback to the DCIM tool. Because rack power consumption varies based on equipment within the rack and its load, each rack should be equipped with a PDU—two for dual bus environments—capable of monitoring power consumption to the rack PDU, as well as overload-protected receptacle groups and, where required, at the receptacle level.
These systems can provide PDU, branch-level, and receptacle-level monitoring of voltage, kW, current, and kWh, giving the most direct measurement of power consumption and supporting higher data center efficiency and availability.
2. Monitoring rack conditions. With increasing densities, a single rack can now support the same computing capacity that used to require an entire room. Visibility into conditions in racks can help prevent many of the most common threats to rack-based equipment, such as accidental or malicious tampering.
If specified and installed correctly, rack monitoring units can be configured to trigger alarms when rack doors are opened, water or smoke is detected, or temperature or humidity thresholds are exceeded. Many units are equipped with cameras to capture video of the event. These “eyes inside the rack” can be connected to central monitoring systems where environmental data can be integrated with power data from rack PDUs, while providing local notification by activating a beacon or other alarm if problems are detected. They should always be deployed in high-density racks and racks containing business-critical equipment. When these sensors and cameras are integrated into the DCIM tool, any alarm or event can be easily diagnosed and treated by the facilities and IT managers responsible for the event.
3. Monitoring energy efficiency. Automating collection and analysis of data from power systems can help reduce energy consumption while increasing IT productivity. Energy efficiency monitoring can track total data center consumption, automatically calculate and analyze PUE, and optimize the use of alternative energy sources.
Using real-time data from the many support systems, the monitoring system can track and trend total infrastructure power use and compare this to capacities. For example, for UPS power output, determine when UPS units are running at peak efficiency, and report Level 1 (basic) PUE. Monitoring at the room or row PDU provides the ability to more efficiently load power supplies, dynamically manage cooling, and automatically calculate Level 2 (intermediate) PUE. Panel board monitoring provides visibility into power consumption by non-IT systems, including lighting and generators, to ensure efficient use of those systems.
Finally, rack-level monitoring provides the most accurate picture of IT equipment power consumption and can support Level 3 (advanced) PUE reporting. The ability to automate data collection, and consolidate and analyze data related to efficiency—which is one of the primary jobs of the DCIM tool—is essential to data center optimization. These functions free up data center staff to focus on strategic IT issues.
4. Monitoring batteries. Batteries have long been known as a weak link in the power chain. To prevent data loss and increase uptime, most data centers specify and require a dedicated battery monitoring system, which should be installed when the data center goes online. Using a predictive battery monitoring method can provide early notification of potential battery failure. The best practice is to implement a monitoring system that connects to and tracks the health of each battery within a string.
The most effective systems continuously track all battery parameters, including internal resistance, using a dc test current to ensure measurement accuracy and repeatability. Supported by a well-defined process for preventive maintenance and replacement, monitoring batteries can significantly reduce the risk of dropped loads due to battery failure, optimize battery life, and improve safety. The use of a battery monitoring system has an excellent payback period but, more importantly, provides peace of mind to the data center operator that the power system will react and support a load during a power outage.
Using intelligent controls
Availability—ensuring IT systems are available when needed—traditionally has been the dominant driver in data center operation. As long as IT systems didn’t go down, the data center was doing its job. However, as organizations struggle to address the rapid pace of technology change and seek out new ways to cut costs—including strategies for improving energy efficiency—a new consideration has emerged: flexibility. Achieving high availability remains the No. 1 priority. And increasing flexibility in how IT systems and data centers are operated and managed increasingly is seen as the best way to achieve that availability while maximizing opportunities to improve energy and operational efficiencies.
Previous generations of infrastructure systems were unable to adjust to dynamic load variations. UPS systems, meanwhile, operated most efficiently at full load, but full-load operation is the exception rather than the norm. The lack of flexibility in power systems led to inherent energy inefficiency.
Integrating intelligent control capabilities into data center power systems can help minimize the amount of energy being lost by enabling data center managers to tailor the performance of the UPS system to the specific efficiency and availability requirements of the site. When proper application of these capabilities is followed, it can enable double-conversion on-line UPS systems to operate at efficiency levels up to 97% while the inverter remains in operation. Advanced UPS systems can employ an energy optimization mode that powers the load in reverse-transfer mode while keeping the UPS inverter at the ready to pick up the load in the event of a utility disturbance. An even higher degree of energy optimization provides power conditioning as well. The result is increased efficiency without compromising availability.
Three key actions that data center managers can take with intelligent control capabilities integrated into power systems include:
1. Energy optimization. An energy optimization, high-efficiency configuration is a unique operating mode that switches the UPS to static bypass during normal operation, increasing efficiency to up to 97% at full load. In this mode, the inverter is left idling while connected to the UPS output, enabling a seamless transfer to double-conversion mode in the event of a utility disturbance, at which time the UPS automatically switches back to double-conversion mode. To ensure the battery is fully charged at all times, the rectifier remains running while the unit is in this mode.
2. Intelligent paralleling. An intelligent paralleling mode increases efficiency by hibernating redundant modules. It places one or more paralleled modules into standby operation when the number of redundant modules rises above the user-specified threshold. When the number of redundant modules falls below that threshold, one or more modules are placed back into normal operation automatically until the number of redundant modules again is greater than the user-specified threshold.
3. Energy conservation testing. An energy conservation testing mode provides an economical and reliable method of unit burn-in and testing that eliminates the need for expensive load bank testing. While in this mode, the inverter is paralleled with the bypass static switch, and power is sent through the rectifier, inverter, and back to the bypass. Data center managers can load test an individual UPS in 1+N configuration, with no risk of tripping the upstream breaker feeding other units protecting the critical load.
UPS systems now include digital controls that can alter and optimize UPS performance. They automatically calibrate the system and ensure the UPS is working properly. They also ensure that the UPS switches between traditional operation and bypass during overloads, protecting UPS systems and the overall power infrastructure.
These controls also enable more efficient operation through energy optimization and intelligent paralleling features. Energy optimization mode increases UPS efficiency by powering the IT load from the bypass path while providing some power conditioning. Energy optimization mode can improve UPS efficiency by as much as 5% but also introduces the possibility of compromising total power protection. This risk can be mitigated when the controls are designed to keep the UPS inverter hot while the system is in energy optimization mode, allowing faster response to utility power disturbances.
Intelligent paralleling manages the load across multiple UPS modules and can automatically deactivate modules that are not required to support the load, while still ensuring that the system is providing adequate redundancy. This can improve system efficiency by up to 6% without sacrificing protection.
It is important that data center managers do not compromise on distribution and server power supply performance. Attention to distribution transformer and server power supply efficiencies can also reap huge benefits in improving PUE. Deploying TP1-rated distribution transformers can improve partial load efficiency by as much as 0.5%. High-efficiency server power supplies, while initially more expensive, can pay for themselves over their useful life.
The challenge for data center managers is to maintain or improve availability in increasingly dense computing environments while reducing costs and increasing efficiency. Applying intelligent control and monitoring capabilities to data center power systems can provide balanced power delivery to better match performance to demand and improve energy savings and PUE. When properly integrated and used, they can help deliver substantial efficiency gains without compromising availability requirements.
- Panfil is vice president and general manager of Emerson Network Power’s Liebert AC Power business, where he leads global market and product development. He serves on AFCOM’s Data Center Institute (DCI) Board. He is a frequent presenter at industry trade shows and conferences, and provides regular expert commentary in leading IT, facilities, and engineering media outlets.