Sergey Nivens - Fotolia
Data center facilities management -- from the maintenance of uninterruptable power supply systems to gauging power efficiency -- is an evolving space for IT teams. For example, older metrics, such as power usage effectiveness, are no longer enough to understand data center efficiency, and increasing power density can lead to more destructive cooling failures.
All of this means it's important to adjust your data center facilities management strategy to adapt to these changing standards, and to keep your data center efficient and cost-friendly. Here are four SearchDataCenter tips from 2016 to help you get started.
Update your data center's battery technology
As data centers become smarter and more efficient, more mature uninterruptable power supply (UPS) systems become a higher priority. Today, there's a demand for UPS systems that have smaller ecological footprints, run cooler and have capabilities to better track data such as lifespan, battery health and maintenance requirements.
Lithium-ion (li-ion) batteries have become a strong contender for battery power in the data center due to their smooth integration with advanced UPS systems. Additionally, li-ion batteries provide more energy and power density, allow for smaller UPS systems that can be placed more flexibly within the data center and are more resilient to higher temperatures than lead-acid batteries.
When choosing a li-ion battery, keep in mind that they aren't hot-swappable, and they aren't necessarily direct replacements for lead-acid batteries. Li-ion batteries tend to cost more than traditional batteries and may have a complex battery management system that adds to the cost, as well.
In 2017, you'll likely see an increase in the availability of li-ion technology. Currently, UPS vendors such as Methode Electronics and Schneider Electric offer li-ion options for several UPS products.
Recognize and prevent data center hot spots
The increased power density of today's data centers can do wonders for efficiency and budget, but also comes with the risk of cooling failures that can cause disastrous server crashes. Hot spots can occur if you move or add equipment without considering the data center's cooling capacities, and can easily go unrecognized until the issue bubbles to the surface in damaging ways.
To find data center hot spots easily and at a low cost, mount temperature-indicating blanking panels -- colored, heat-sensitive strips that indicate inlet air temperatures -- near the top, middle and bottom of each rack, or in front of the most vulnerable hardware if your budget is tight.
Temperature and humidity probes are another option -- either as an add-on to smart rack power distribution units, wireless devices or part of data center infrastructure management (DCIM) tool -- that can offer real-time graphical displays of data center temperature. To simulate an installation and test out cooling capabilities, combine computational fluid dynamics (CFD) air flow modeling with the probes' readouts. In fact, the best way to prevent cooling failures is with CFD, by creating a 3D model of the data center.
Measure efficiency accurately with new metrics
The Green Grid introduced the power usage effectiveness (PUE) metric nearly 10 years ago. Since then, data center technology has evolved significantly, rendering the PUE less sufficient in accurately measuring data center efficiency. To get a true sense of efficiency, IT teams require detailed metrics that include power and cooling infrastructure as well as computing systems. Fortunately, data center teams today can choose from a variety of metrics.
Data center energy productivity (DCeP) measures the amount of useful work the data center produces per watt of consumed energy -- with useful work defined by the user. For example, a retail company may define useful work as the number of sales. The goal of the DCeP metric is to minimize energy consumption and maximize useful work.
Another new metric to gauge data center efficiency is Performance Indicator (PI) -- which Green Grid released in June 2016. This metric accounts for PUE, IT thermal compliance and IT thermal resistance. Use PI to understand how reliably hardware is being cooled, the facility's efficiency and how one factor affects the other. PUE still remains as the basis for efficiency metrics, but new metrics such as PI can enhance data center facilities management in more nuanced ways.
Increase efficiency with data center cable management
Another way some IT teams unintentionally waste resources when it comes to data center facilities management is with unmanaged cabling. Poorly placed cabling can create heat and stop the flow of cooling air in the under-floor plenum, driving up costs and causing potential downtime.
To start the cleanup process, use a DCIM tool to take full inventory, from servers and switches to network hubs and supplementary equipment. With the DCIM tool, perform simulations on potential layouts to determine the most effective design, as well as analyze your current wiring layout to identify which cables are no longer live. This can help you to clean up without creating any accidental downtime.
Another approach is to implement fully structured cabling, which can be either under floor or above floor. An above-floor, wiring tray-based method is best, since under-floor cabling can limit access to the cables. Label and color the cable runs to easily identify them, and cut each cable to the required length without looping them.
The final and most expensive approach is to build a data center using the tiering model, which separates facilities into four categories based on their availability.
Upgrade power, cooling and more to increase your facility's lifespan
Maximize efficiency with three-phrase circuits
Understand data center metrics with this guide