Data center operators are under increasing pressure to make their facilities more energy efficient. An increasing number of large cities and countries are mandating a high level of DC greenness for new DCs and considering extending the exacting requirements to existing ones, and energy prices are high and appear to stay that way going forward. Exacerbating the problem are the global chip shortage and supply chain glitches, which delay the initiation and completion of projects that involve replacing older assets with more energy efficient ones.

Enterprise DC owners employ a variety of measures to make their facilities a deeper shade of green. These include offloading data to the cloud, switching to flash storage and solid state drives, using variable frequency drives and energy efficient motors in cooling systems, and virtualizing or consolidating DCs. Others move some of their applications offsite to take advantage of providers’ typically more energy efficient infrastructure, or employ renewable energy.

Variety of methods
Some of these measures, however, may not be available to some operators for economic, enterprise policy, regulatory or other reasons. This is especially so for those with existing DCs. Thankfully, other methods of reducing energy usage are available and DC operators would do well to consider them seriously.

A good starting point is cooling. Many operators use the nameplate power ratings on their servers and other assets to size their cooling systems and set a baseline temperature for their facilities. Using such peak load numbers, however, results in overcooling as DC computing assets almost never run at peak capacity. It has been estimated that DC energy costs can be lowered by up to four percent for every degree of upward change in baseline temperature. Given that DC cooling accounts for a sizeable chunk of total energy consumption, any percentage drop is significant not just in the financial sense but environmental as well.

Remote power management is another way of improving energy efficiency. By using intelligent PDU with the right capabilities, DC operators can, from afar and over a secure network, lower their energy costs by simply powering off servers and other equipment that do not need to run 24/7. When used with intelligent racks, such PDUs can also measure power consumption by each rack server.

DC insights
Using such insightful data, they can monitor and optimally configure the available power capacity for each rack, maximize rack space, switch and control power outlets, calculate PUE accurately, identify degrading power supply assets, and better plan power capacity and cooling infrastructure with a view to conserving energy.

Obviously, the ability to measure and monitor facility and rack temperature, asset power consumption, and other parameters needs to be present before the abovementioned measures can be undertaken. Sensors, smart rack controllers and other tools are readily available for remote and accurate measurement and management of power, temperature, humidity and other parameters. When used in concert with DCIM and BCM systems, they can enable DC operators to make their facilities perform at a higher level and yet be friendlier to the environment.

Moving up
Another way of lowering energy usage is to make optimal use of the vertical space in DCs instead of investing in additional racks and having to cool them. This can be done by simply tidying up the space within each rack, creating more space within racks by repositioning patch panels, network switches and other gear outside the racks, and switching to taller racks.

Be continually employing these methods to lower energy usage in their facilities, DC operators can, over time, significantly cut the amount they spend and contribute to long-term sustainability. Raritan offers a suite of hardware and software solutions that can help enterprises achieve a high level of energy efficiency. These solutions include network-connected PDUs, intelligent rack controllers, sensors, and DC energy management and power monitoring software.