Data Center Infrastructure Management( DCIM ) software has become a critical piece of a reliable data center. DCIM software is used to monitor critical devices, whose uptime is essential for the data center. It is also used to monitor space, power availability and cooling capacity of a data center. Available capacity of power, space and cooling are three pillars on which a data center functions.

In a small data center the essential parameters such as cooling, power and space may be monitored by staff. However, with larger data centers, whether captive or colo, it is impossible to manage the operations maintaining SLAs such as uptime and PUE metrics without relying on DCIM software.

There are two critical aspects that need to be considered when choosing DCIM software for medium to large data centers : scalability and reliability. Most enterprise software will have similar requirements on scalability and reliability.

A large data center will have hundreds of racks and electrical devices such as UPS, PDUs, panels etc. A scalable DCIM software should be able to add & configure these devices as quickly as possible. Once the devices have been configured, DCIM software should be able to monitor hundreds of these devices without adding to the latency. Hence, the DCIM software must be inherently scalable to configure large number of devices and have underlying thread support to utilize multiple cores of a server and be able to scale horizontally and vertically. Without this it will choke after certain number of devices are added.

The aspect that goes hand in glove with scalability is reliability. DCIM software must have 99.99 % availability. The software must gracefully deal with exceptional conditions arising out of device failure or bad data provided by malfunctioning devices. It must have robust error handling capabilities and its logs must be able to point to cause of any failure.

In summary, due consideration should be given to scalability and reliability aspects of DCIM software especially when it is used for a medium to large data center. Without those capabilities it may not give the ROI as anticipated.