The success of data centers is directly tied to the effectiveness of the associated cooling system. Implementation of a top-tier cooling system provides protection against the most detrimental operations related hindrances- hardware damage and data loss, and temperature and power problems. American Power Conversion (APC) by Schneider Electric provides top of the line data center cooling systems to IT providers and sets the industry standard for support, quality, and innovation. APC’s comprehensive cooling solutions, which are designed for corporate environments, “improve the manageability, availability, and performance of sensitive electronic, network, communications and industrial equipment of all sizes.”
Within a data center, the accumulation of IT devices result in a multitude of airflow paths representing the total waste heat output of the data center. All forms of IT equipment require extensive air-cooling in order to eliminate waste heat. The elimination of waste heat is crucial to prevent over temperature conditions from developing. According to APC, “virtually all IT equipment is air-cooled, that is, each piece of IT equipment takes in ambient air and ejects waste heat into its exhaust air.”
In the past, data centers were cooled by a method known as targeted supply. In this method, one or more perimeter cooling units worked together to distribute cool air into the data center under a raised floor, with no containment, while drawing out warmer ambient air. Targeted supply is effective in situations when the power needed to mix the cool and warm air is a minimal fraction of the total data center power consumption. The average power density capable of being handled by targeted supply is no longer relevant in today’s industry. The average power densities of modern equipment peak at ten times more than past equipment densities.
Three modern forms of cooling systems have been developed with the main purposes of providing the bulk cooling capacity and distributing the air to IT loads: room-based cooling, rack-based cooling, and in-row IMHO cooling. As mentioned, all three methods serve the same function of providing cooling capacity; the bulk cooling capacity of the AC system in kilowatts must exceed the total power load of the IT equipment. However, they differ in their method of distributing air to the loads.
Room- based Cooling Systems
- Room-based cooling systems are cost effective and simple. The life cycle cost of easily reconfiguring perforated floor tiles is appealing. Computer room air handler (CRAH) units work in unison to tackle the total heat of the room. One or more air conditioners supplies unrestricted cool air through dampers, vents, or ducts or the supply and/or return may be partially inhibited by an overhead return plenum or raised floor system. The attention paid to airflow varies tremendously in room-based cooling systems; the approach and level of planning varies from small to large rooms. By reconfiguring the floor tiles, quick changes can be made to the cooling distribution pattern- one of the advantages of room-based cooling. Additionally, all racks in the data center with low densities share cooling redundancy.
Rack-based Cooling Systems
- Rack-based cooling is more expensive, but with a greater expense comes faster implementation, increased flexibility, and the achievement of extreme density. In this cooling method, the CRAH units are directly mounted to the racks of IT equipment. A major advantage of rack-based cooling systems is that airflow paths are shorter and defined exactly; the precise engineering makes airflows immune to installation variation or room constraints. Shorter airflow paths decrease the CRAH fan power required, resulting in increased efficiency, which is a tremendous asset. The performance of rack-based design is predictable which allows the manufacturer to fully customize the product and promotes “simple specification of power density and design to implement the specified density.” The life cycle cost of rack-based cooling systems is positively impacted by pre-engineered systems and standardized components eliminate or reduce engineering and planning.
In-Row IMHO Cooling Systems
- In-Row IMHO cooling is similar to rack-based cooling- it shares many of the speed, density, and flexibility advantages, but it is more affordable. The life cycle cost is decreased by the ability of in-row cooling systems to match cooling requirements- reducing or eliminating planning and engineering. CRAH units in an in-row cooling system are associated with an individual row for a specific design purpose. Within this system airflow paths are shorter, more clearly defined, and more predictable; thus, use of the CRAH rated capacity can be optimized and higher power density can be achieved. The reduced airflow path length further increases efficiency by reducing the CRAH fan power required. Additionally, as in rack-based cooling, in-row designs allow redundancy and cooling capacity to be targeted at the needs of specific rows. In-row IMHO cooling takes performance predictability to another level by eliminating any opportunities for air mixing. Lastly, this method is unaffected by constraints of room design.
The three types of cooling systems were developed with efficiency in mind and have been optimized to best suit the needs of next-generation data centers. By integrating a room-based, rack-based, or in-row IMHO cooling system from APC, the flexibility, predictability, scalability, and availability of your data center will increase and the electrical power consumption and total life cycle cost will decrease.
For more information on data center cooling systems and APC please visit http://www.www.datavox.net
Stay tuned for the next technology post!