You wouldn’t want an annual physical without the doctor checking your heart rate, measuring your blood pressure, and in some cases running an EKG and stress test. Just as your heart is a critical component of your overall health, the power-system infrastructure is critical to the smooth functioning of a modern data center. Most conservative estimates place 50–70 percent of data center asset value in its power systems—think circuit breakers, transformers, electrical feeders, UPSs, power-distribution units, generators and so on. Data center operators wake up in a cold sweat from the nightmare of an unplanned outage resulting from power-related problems. Both from total-cost-of-ownership and operational-expense perspectives, the “heart” of the data center is undoubtedly its power system.
Figure 1. Typical data center asset breakdown.
Focusing attention solely on IT server-management tools is like checking your sugar levels but skipping that heart check during a physical. Those unglamorous, humble gray boxes that keep the data center humming and sales churning certainly deserve some TLC! That’s why a dedicated energy- and power-management system (EPMS) should be an indispensable component of your facility strategy. Just as building-management systems (BMSs) bring great visibility to cooling and facility mechanical systems, power-management systems bring data visibility and analytics to valuable electrical assets. Without a well-tuned EPMS, a data center infrastructure management (DCIM) suite risks having serious blind spots.
Let’s focus on five main benefits that EPMSs offer.
1. Visibility of Electrical Data
Data center operators need 24x7 visibility to critical electrical data. Let’s consider some concrete examples.
a. Granular electrical data: Data from power and energy meters help answer operational questions such as “How much load is that circuit carrying? Do I risk tripping if I add more server load? Is that transformer overloaded?” Timely access to such data can help optimize electrical capacity, so you “sweat” your assets but avoid compromising your redundancy.
b. Functional-operator screens: Quickly navigating equipment-status information to look at critical power infrastructure such as transformers, UPSs and breakers helps operators actively identify problem areas that could potentially hurt uptime.
c. Timely electrical alarms: Identifying overload conditions and unexpected circuit-breaker trips through appropriately set alarm thresholds can help data center operators intervene before a minor problem escalates to a crisis. An EPMS brings faster alarm performance and accurate time stamping, both of which help determine effectiveness in responding to alarm conditions.
d. Maintenance-related electrical data: Analyzing circuit-loading history can inform a more precisely targeted maintenance plan. For example, analysis of interrupted loads and operations may suggest a circuit breaker needs maintenance earlier than normally scheduled. Conversely, deferring scheduled maintenance may be equally warranted because a piece of equipment has been lightly loaded or has operated less than originally anticipated. Such targeted maintenance plans can save considerable operational cost in a large data center.
2. Improved Electrical-Safety Procedures
Operating and maintaining critical power equipment in a modern data center requires careful attention to safety. Your operators need timely access to electrical data before working on live equipment to comply with codes and standards such as the arc-flash safety standard required by NFPA 70E. Equipment status screens help reduce risk of catastrophic injury through inadvertent operator error. After all, human error continues to be a leading cause of data center outages (see Figure 5).
Software can also help in the safe operation of equipment through “virtual” tagging of electrical circuit breakers and switchgear. Additionally, role-based security features can enhance security and safety of remotely operated equipment by ensuring that only the right personnel have access.
3. Root-Cause Analysis and Event Forensics
Power grids are under stress today because of natural weather events and stretched capacity, so outages are not exactly uncommon. And let’s face it: in a complex facility such as a data center, things can go wrong. Figuring out precisely what happened when and, more importantly, the exact cascading sequence of events is crucial to ensuring it doesn’t happen again. That’s where the extra horsepower of a power-management system is handy insurance, especially in post-event crisis mode. Here are three areas where the event data captured by EPMSs can be invaluable:
a. Electrical alarms: Tracking voltage disturbances such as sag/swell events and circuit-breaker trips is imperative to assembling evidence of incorrect operation. Whether it’s to improve internal operations or to negotiate power-quality contracts with the utility, ascertaining facts with hard evidence is essential.
b. High-accuracy sequence-of-events reporting: In the high-speed world of electrical operations, approximating to seconds or minutes doesn’t cut it. Sub-cycle accuracy (1- to 10-millisecond resolution) is necessary to isolate the culprits in an unexpected outage: mis-coordinated relays, stuck breakers and so on.
c. Power-quality waveform analysis: An underappreciated fact of modern electrical systems is that an unintended consequence of increasing energy efficiency and reliability through UPSs, switched drives and so on is that it actually contributes to reducing power quality. As the cost of power-quality-related outages continues to rise, EPMSs can help analyze electrical faults, harmonics and other electrical disturbances through waveform-analysis tools. Quickly isolating the direction of the disturbance’s origin through disturbance-direction detection can save valuable time after a power-system event.
4. Energy Accounting and Cost Allocation
Although data centers are understandably obsessive about reliability and uptime, they are increasingly being tasked with energy-efficiency programs as well. Data centers are also rapidly adding metering at highly granular levels to accurately bill colocation customers. EPMSs can facilitate colocation billing by accurately allocating net energy costs including ancillary charges for peak demand and power-factor charges. In many cases, data centers can recover a larger proportion of their energy costs by billing customers more accurately.
In addition to externally billed IT customers, cost-accounting processes at many data centers now require internal billing at the departmental or cost-center level. EPMS software centralizes much of metering to output the data required for these applications.
5. Sustainability and Operational Metrics
Executives now expect data center personnel to have operational metrics available at their fingertips. Here are four areas where an EPMS can enhance both the quantity and quality of available metrics:
a. Capacity management: By analyzing circuit loading at various points in the electrical infrastructure, operators can assess whether they can safely add IT load in a particular area without compromising design redundancy.
b. Operational performance: Automating PUE calculations per applicable standards (such as Green Grid recommendations) can make life easier for operators by eliminating spreadsheets.
c. Energy reporting: Energy-consumption reporting at a granular level is driven by the need to verify conservation measures and by compliance with energy-performance standards, such as ISO50001.
d. Sustainability reporting: Sustainability is no longer an afterthought in boardrooms. Adoption of CO2-emission reporting, targets and mandates are increasing in the data center space as well and require a solid metering-system infrastructure.
Paying closer attention to your data center’s electrical power infrastructure through a dedicated EPMS is well worth the time, money and effort. From an ROI perspective, this investment can be likened to “insurance coverage” where investing a fraction of the capital cost is simply the smart thing to do.
Yes, building-management systems can certainly integrate power equipment to some extent, but that’s not what they are built for. BMSs are brilliant at managing mechanical systems, but relying on them completely for your power infrastructure would be underestimating the value of the heart of your data center.
And don’t worry, integrating with other critical facility systems is not the horror story it once was. Components of a DCIM suite are increasingly designed with integration in mind, and building management, data center operations and power management all play nicely on the same playground.
Make sure the power systems in your data center are well monitored and tuned—your data center needs its heart in the right place.
About the Author
Ram Kaushik has 25+ years’ experience in various technical and business roles in engineering, software development and business, the energy/power sectors, software consulting, and technology startups. He is currently U.S Offer Manager for Schneider Electric’s Energy & Power Management Solutions, managing a portfolio of hardware, software and services offers targeting a variety of segments including data centers, health care, industry, commercial facilities and utilities. He has presented and taught extensively at seminars and training events, both at Schneider Electric and other technology forums. Having a passion for technology and software, Ram is an incurable software tinkerer and programmer. He holds an MBA from Vanderbilt University and an MS in electrical engineering from Rensselaer Polytechnic Institute.