Modern IT hardware has increased in density over the past decade. Individual racks now consume more than 30,000 watts of energy.

IT technology has increased in density over the past decade to the point where individual racks can now require more than 30,000 watts of energy. In 2007, an article by John Niemann, product line manager at APC-MGE, reported that new server technologies were driving power densities in excess of 30 kilowatts per cabinet. More recently, a Feb. 24 blog posted on the Kratos Network Solutions website noted that new servers have pushed the power demands of some individual rack levels past 30,000 watts.

As density increases, so does the heat dissipated into the data center. As a result, the data center cooling infrastructure must be capable of supplying sufficient airflow to meet new demands when and where cooling is needed.

Raised access floors have long been the standard for meeting the cooling needs of the industry, while providing an effective method for distributing a large numbers of data and power cables.

According to Scott Alwine, marketing manager for Tate Access Floors, raised access floors have long been the standard for meeting the cooling needs of the data center industry, while providing an effective method for distributing large numbers of data and power cables and other services. “The system easily adapts to technological and client changes, while offering a solution for distributing water and other liquid cooling agents to row- and rack-based equipment, minimizing the threat posed by water leaks or condensation due to system failure,” Alwine said. “Raised floors also provide a platform for future scalability and the flexibility to incorporate energy-efficient design opportunities that support increasing heat loads.”

The flexibility a raised floor system offers has become even more important as data centers continue to evolve, giving rise to a variety of cooling challenges. Until recently, load profiles per rack were relatively stable in terms of energy consumption and heat production. Changes in electrical and airflow distribution were limited to moves, adds, or changes in the data center environment. Today, however, demands for both energy and cooling increase from one year to the next, even as load profiles vary from rack to rack and fluctuate within a given rack from one minute to the next. A white paper (“Guidelines for Specification of Data Center Power Density”) published by American Power Conversion references the fact that “… IT equipment is constantly being refreshed, which means the power consumption of particular racks is subject to change over time.”


These increases create incentives for the industry to improve both efficiency and capacity. The growth in popularity of cloud computing is one such improvement. In fact, a 2010 presentation by Scott & Scott, LLP, included an estimate by Gartner Research that the cloud market would reach $150 billion by 2013. IT hardware can now be used at a much higher level, because it exists as a component of the cloud. The scalable cloud allows for computing resources to be brought online as required by demand to realize the efficient and complete utilization of the hardware.

As a result, heat loads within data centers now vary over the short and long term and on a rack-by-rack basis, impacting airflow requirements from one rack to another. To further complicate data center cooling requirements, individual loads within the rack fluctuate throughout the day due to the processing demands of IT hardware.

Matching cooling to the load becomes an increasingly important factor in the cloud environment.

Given these circumstances, the key to success in the data center is the ability to control the amount of air entering the rack and directly match it to airflow requirements of the IT hardware at any given moment in time. In the typical airflow path of data centers utilizing a raised floor plenum, computer room air handling (CRAH) units installed on the floor feed air into the raised floor plenum where it is pressurized and then forced through perforated panels. These interchangeable, perforated panels, along with high airflow grates, offer flexibility through the use of manual dampers, but not to the degree that variable heat loads now require.

To meet the challenges that these ever-changing loads create, new technologies are emerging for raised floor designs, including local and dynamic airflow delivery technology and high total air capture airflow products. These innovative devices allow for sufficient airflow delivery to the IT load while ensuring the delivery of the precise amount of air to the IT equipment, which also ensures reliability and lower energy costs.


Historically, perforated panels similar to the solid panels used in raised floor systems have provided cold air to IT equipment. In most cases these panels are perforated to provide an open area of approximately 25 percent, and airflow delivery through the panels is a function of the differential pressure across the panel.

As demands for cooling increased, new panels emerged-grates that offer a far greater open area. Some of the leading designs feature maximum open areas of up to 56 percent. With this increased open area come higher flow rates vs. static pressure-airflow in excess of 3,000 cubic feet per minute (cfm), nearly three times that realized with a 25 percent perforated panel.

Variable speed fans increase airflow to loads in response to increased demands and then throttle back down.

However, not all of this air reaches its intended target-the rack and the equipment it supports. According to Bill Reynolds, director of technical services for Tate Access Floors, “Although part of the airstream does enter the IT rack, a significant portion of the vertical air plume that emerges from the panel bypasses the equipment entirely, mixes with hot exhaust air and either returns to the air handling units or is ingested at the rack intake.”

The amount of air entering the rack makes up the panel’s total capture rate, defined as the total airflow captured at the IT load divided by the total airflow from the panel. “In effect, the perforated panel delivers to the equipment only a portion of the air that flows through it at any given time-typically below 50 percent,” Reynolds explained. “The air that is not delivered to the face of the rack becomes bypass air and results in wasted energy and a reduced cooling capacity.”


Recent improvements to the grates that provide airflow from the underfloor plenum are addressing these compromised capture rates. These improved grates move beyond slight increases in open areas to actually direct airflow to the racks. 

Interchangeable perforated panels and high airflow grates ensure flexibility.

Directional grates feature angular vanes built into the structural component that direct the plume of air evenly to the entire face of the IT rack, improving the total volume of air that is delivered to the rack. According to a Tate white paper titled, “Optimizing Capacity and Efficiency in a Diverse and Variable Load Environment,” directional grates further boost the maximum capacity and energy efficiency for a given grate by providing a total capture rate as high as 93 percent-nearly eliminating bypass air without the need for aisle containment. And by significantly reducing bypass air, operating expenses for cooling are reduced by as much as 40 percent, along with the capital expenditure on cooling equipment in new facilities. For example, delivering air more effectively in an angular plume to the rack may reduce the number of CRAH units required to adequately cool a data center.


Although the use of directional grates improves the distribution of air to all points of the rack face, minimizing bypass air and improving efficiencies in the process, these grates do little to address the issues posed by diverse and variable loads. Data centers today rarely present a homogeneous load profile, the result of rack load diversity. While some racks in the average data center have yet to be deployed, other racks may be fully deployed but with equipment that requires very little airflow. Still other racks may support high-density equipment that requires 10 to 20 times the airflow of the rack directly beside it. 

High airflow directional grates improve rack kW capacity

In the past, dampers have been employed to meet such load diversity. Different airflow panel designs based on the airflow requirements of the rack have been the alternative. Unfortunately, airflow panels cannot always provide the precise amount of air when it is needed nor can they account for cooling demands that vary during the course of a day.

Although dampers are able to address diverse rack demands, they must be adjusted manually to restrict or increase airflow as demands change. Additionally, the process of determining the airflow exiting an airflow panel, adjusting the damper to meet the rack requirements, and then moving to other panels in the system can result in changes in the static pressure throughout the site, requiring additional interaction before balance is achieved. Then, any additional moves, adds, or changes at the rack level may require this tedious process to be repeated throughout the environment.

 “Further complicating the ability to achieve proper airflow balance is the variability of the load in any one rack,” said Reynolds. “As equipment responds to activity, heat load profiles can swing dramatically, by rack and by equipment within a single rack. Therefore, if manual adjustments to airflow are made, they need to account for the maximum requirements during periods of peak computing demand, which may make up only a short period of time throughout the day.” 

Variable-air-volume dampers help address variable load.

One of the newest solutions to the challenge of diverse and variable demand at the rack level is a variable-air-volume (VAV) device installed below each airflow panel and sized to handle the volume of air required by the rack the panel serves. This VAV device is electronically controlled, effectively adjusting the amount of air passing through an individual panel to meet the rack’s specific cooling needs. Sensors mounted to the front of the rack control the VAV damper and, consequently, the airflow amount, maintaining the proper inlet air temperature on a rack-by-rack basis. This flexibility helps data center managers implement strategies for virtualization, cloud computing, and servers with idle standby modes while saving energy.

In effect, this VAV device adapts technologies readily found in commercial office applications. Each panel and rack tandem can be thought of as an individual zone. The VAV device measures the incoming air temperature at the face of the rack and adjusts the airflow to ensure that the temperature at the rack’s face doesn’t exceed the maximum allowable setpoint.

Furthermore, when coupled with a directional grate, the system nearly eliminates bypass air and accounts for local temperature fluctuations. The system also provides indirect feedback to the air-handling equipment regarding the required airflow through pressure transducers installed in the raised floor plenum. As the VAV damper closes, the underfloor static pressure increases, indicating to the air-handling unit to slow its fan speed, saving energy. This ability to reduce fan energy reduces energy consumption in the data center, particularly during periods of non-peak computing demand, and improved power usage effectiveness (PUE).


It is not unusual for individual racks within a data center to exceed the capacities of an individual airflow panel. To meet these demands for spot cooling or to serve areas of the data center where airflow is low due to insufficient raised floor height or other underfloor restrictions, data centers are turning to fan-assist modules.

Thermostatical controls adjust the amount of air passing through an individual panel to meet the rack’s specific cooling needs and ensure precise temperature control at the inlet of each rack.

The fan-assist module effectively manages the cooling requirements for dense server racks and blades by providing a blast of cooling through an individual airflow panel. Using temperature sensors mounted to the front of the rack, the module automatically turns on when conditions require additional cooling.

A fan-assist module also increases the airflow in low-pressure zones by monitoring the entering air temperature at the rack inlet, thereby ensuring adequate airflow.

Fan assist modules effectively manage cooling requirements for dense server racks and blades.

Equipped with a variable-speed fan drive, a fan-assist module can be throttled up or down based on heat-load requirements. This makes the module the ideal solution for cooling the toughest hot spots in a data center. In addition, it can be installed with a directional grate to control bypass air as it throttles up or down. And, when combined with computer room air conditioners (CRAC) and CRAHs that use variable-frequency drives or electronically commutated (EC) fans, the fan-assist module contributes to savings gained through the ability to deliver higher volumes of air at lower static pressures.

Each of these emerging technologies-directional grates, VAV dampers, and fan-assist modules-individually and in combination, effectively address the challenges of increased cooling demands and diverse and variable loads. These capabilities will likely become more important as cloud computing grows in popularity, pooling the resources of multiple servers throughout the data center and resulting in dramatic swings in the heat load profile. The good news is, as they meet these new challenges, these same solutions also provide tools to help data center operators achieve energy efficiency goals while reducing operating costs and improving PUE.