Redundancy Ensures Data Center System Availability if Components Fail
Redundancy ensures system availability in the event of component failure. Two types of redundancy — active/standby and active/active — are of greatest interest.
An active/standby approach does not actively participate within the system during normal operation, while an active/active system does. An active/active approach is considered superior in terms of performance and resiliency because there is little, if any, downtime. Modular systems are an example of an active/active approach, which also includes systems such as modular boilers, chillers, dry coolers, and fan-wall technology.
These systems satisfy changes in loads due to increased cooling and power requirements, and they eliminate interruptions due to component failures. Modular units can work well in retrofits due to their smaller space requirements.
Redundant power and fuel sources also are important considerations. Examples include the use of dual gas/oil heating systems, direct-expansion split systems, and variable-refrigerant air conditioning systems used to supplement water-cooled air conditioning, such as perimeter computer room air conditioners and in-row direct expansion units.
Another application of power and fuel redundancy in data centers involves the use of gas-fired microturbines, which can provide a source of routine and emergency power to prevent interruptions. A turbine connected to an absorption chiller also provides a redundant source of heating and cooling.