Hardware25.06.2013

Cooling today’s sizzling hot data centres

By Kevin Dunlap, general manager of cooling solutions, and Neil Rasmussen, senior VP of innovation for Schneider Electric

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. Since a data centre may contain thousands of IT devices, the result is that there are thousands of hot airflow paths within the data centre that together represent the total waste heat output of the data centre; waste heat that must be removed.

With power densities of modern IT equipment pushing peak power density to 20 kW per rack or more, simulation data and experience show traditional cooling, dependent on air mixing, no longer functions effectively.
To address this problem, design approaches exist that focus on room, row, and rack-based cooling. In these approaches, the air conditioning systems are specifically integrated with the room, rows of racks, or individual rack in order to minimise air mixing.

Every data centre air conditioning system has two key functions. The first function of providing bulk cooling capacity is the same for room, row, and rack-based cooling. The major difference lies in how each cooling system performs the second critical function, distribution of air to the loads. Controlling the airflow is therefore the main objective of the different cooling system design approaches.

Room-based cooling

With room-based cooling, the computer room air handler (CRAH) units operate concurrently to address the total heat load of the room. Room-based cooling may consist of one or more air conditioners supplying cool air completely unrestricted by ducts, dampers, vents and more; or the supply and/or return may be partially constrained by a raised floor system or overhead return plenum.

The room-based design is heavily affected by the specific constraints of the room, including the ceiling height, the room shape, obstructions above and under the floor, rack layout, CRAH location, the distribution of power among the IT loads and more. When the supply and return paths are uncontained, the result is that performance prediction and performance uniformity are poor, particularly as power density is increased. Therefore, with traditional designs, complex computer simulations called computational fluid dynamics (CFD) may be required to help understand the design performance of specific installations. Furthermore, alterations such as IT equipment moves, add-ons, and changes may invalidate the performance model and require further analysis and/or testing. In particular, the assurance of CRAH redundancy becomes a very complicated analysis that is difficult to validate.

Another significant shortcoming of uncontained room-based cooling is that in many cases the full rated capacity of the CRAH cannot be utilised. This condition occurs when a significant fraction of the air distribution pathways from the CRAH units bypass the IT loads and return directly to the CRAH. The result is that cooling requirements of the IT layout can exceed the cooling capacity of the CRAH despite the required amount of nameplate capacity.

Row-based cooling

With a row-based configuration, the CRAH units are associated with a row and assumed to be dedicated to a row for design purposes. The CRAH units may be located in between the IT racks or they may be mounted overhead.

Compared with the traditional uncontained room-based cooling, the airflow paths are shorter and more clearly defined. In addition, airflows are much more predictable, all of the rated capacity of the CRAH can be utilised and higher power density can be achieved.

Row-based cooling has a number of side benefits other than cooling performance. The reduction in the airflow path length reduces the CRAH fan power required, increasing efficiency.

This is not a minor benefit, when we consider that, in many lightly loaded data centres, the CRAH fan power losses alone exceed the total IT load power consumption. A row-based design allows cooling capacity and redundancy to be targeted to the actual needs of specific rows. For example, one row of racks can run high-density applications such as blade server, while another row satisfies lower power density applications such as communication enclosures.

For new data centres less than 200 kW, row-based cooling should be specified and can be implemented without a raised floor. For existing data centres row-based cooling should be considered when deploying higher density loads (5kW per rack and above).

Rack-based cooling

With rack-based cooling, the CRAH units are associated with a rack and are assumed to be dedicated to a rack for design purposes. The CRAH units are directly mounted to or within the IT racks. Compared with room-based or row-based cooling, the rack-based airflow paths are even shorter and exactly defined, so that airflows are totally immune to any installation variation or room constraints. All of the rated capacity of the CRAH can be utilised, and the highest power density (up to 50 kW per rack) can be achieved. The reduction in the airflow path length reduces the CRAH fan power required, increasing efficiency. This is not a minor benefit considering that in many lightly loaded data centres the CRAH fan power losses alone exceed the total IT load power consumption.

A rack-based design allows cooling capacity and redundancy to be targeted to the actual needs of specific racks, for example, different power densities for blade servers vs. communication enclosures. Furthermore, N+1 or 2N redundancy can be targeted to specific racks. By contrast, row-based cooling only allows these characteristics to be specified at the row level, and room-based cooling only allows these characteristics to be specified at the room level.

As with row-based cooling, the deterministic geometry of rack-based cooling gives rise to predictable performance that can be completely characterised by the manufacturer. This allows simple specification of power density and design to implement the specified density. Rack-based cooling should be used in all data centre sizes where cooling is required for stand-alone high-density racks. The principal drawback of this approach is that it requires a large number of air conditioning devices and associated piping when compared to the other approaches, particularly at lower power density.

Hybrid cooling

Nothing prevents the room, row, and rack-based cooling from being used together in the same installation. Placing various cooling unit in different locations in the same data centre is considered a hybrid approach. This approach is beneficial to data centres operating with a broad spectrum of rack power densities.

Another effective use of row and rack-based cooling is for density upgrades within an existing low-density room-based design. In this case, small groups of racks within an existing data centre are outfitted with row or rack-based cooling systems. The row or rack cooling equipment effectively isolates the new high-density racks, making them “thermally neutral” to the existing room-based cooling system.

Another example of a hybrid approach is the use of a chimney rack cooling system to capture exhaust air at the rack level and duct it directly back to a room-based cooling system. This system has some of the benefits of a rack-based cooling system but can integrate into an existing or planned room-based cooling system.

To make effective decisions regarding the choice between room, row, or rack-based cooling for new data centres or upgrades, it is essential to relate the performance characteristics of the cooling methods to practical issues that affect the design and operation of real data centres.

Sign up to the MyBroadband newsletter