How robust data center monitoring pays for itself

How robust data center monitoring pays for itself

The cost savings from data center monitoring add up over time.

Data center infrastructure management and other data center solutions have become mainstays in today's IT facilities. The rate at which new data is being created and our indisputable dependency on this data for operational excellence have necessitated streamlined data center management. 

Among these solutions is data center monitoring, which acts as a facility's eyes and ears. Without data center monitoring, DCIM cannot as effectively supply up-to-the-second insight about environmental conditions within a facility. This is but one of the key ways that robust data center monitoring shows return on investment. Other sources of ROI include the following:

Preventing downtime

Downtime is the scourge of the data center, costing nearly $8,000 per minute, on average. To alleviate the lost opportunity and reputational damage that comes from downtime, monitoring solutions collect data about the following conditions:

  1. Temperature
  2. Humidity
  3. Flooding
  4. Power consumption
  5. Security

In theory, any one of the above factors can precipitate downtime. Temperature spikes or hot spots can cause servers to overheat. Too much humidity can corrode electronics, while a complete lack thereof causes static discharge that can damage them. Flooding and excess condensation, likewise, can lead to downtime. Meanwhile, overburdening certain power supply feeds can lead to shorts, and voltage spikes can result in breaker activations. Last but not least, a security breach can knock a facility out of commission. Back in 2014, a Danish ISP experienced downtime after a thief cut a hole in the wall to steal a few networking cards. 

All of these threats sound alarming. However, those that can't be prevented by monitoring (i.e. acts of nature) can be responded to within seconds. The very moment that an allowable threshold is exceeded, alerts are sent to designated staff via SNMP, SMS or email. A swift course of action can then be taken to mitigate potential losses from downtime. 

You can't prevent some causes of downtime, but quick reaction time may save some of your assets.Some causes of downtime are unpreventable, but quick reaction time may save some of your assets.

Conserving resources

"There were about 10 million zombie servers globally at last count."

Downtime may immediately cripple a business, but resource inefficiencies slowly drain money. Cooling, for instance, comprises 40 percent of total data center power consumption. So, just as too high a temperature causes problems in the data center, lower than necessary temperatures may mean that more cooling capacity than necessary is being occupied. Equally important, strategically placed temperature monitors can identify anomalous hotspots in a data center, helping managers figure out where their airflow management may be weak.

If that moisture starts to settle on the ground as condensation, if the ceiling starts to leak or if water begins to accumulate for any other reason, flood-sensor cables along the ground will determine if the floor is dry, damp or completely submerged in water. If necessary, an automatic alert will notify personnel as soon as moisture is detected by the cable. This ensures that leaks are addressed quickly so equipment experiences the minimum amount of damage. 

Finally, power monitors let staff track the amount of energy in use. This makes it possible to identify equipment that is operating inefficiently and pinpoint zombie servers – those that occupy space and energy without actually being in use. According to the Natural Resource Defense Council, there were about 10 million zombie servers globally at last count. This equates to "$30 billion in server capital."

Robust data center monitoring helps data center managers realize these and other savings in their facilities, while helping to prevent downtime. In doing so, it pays for itself, and then some. To learn more, contact Geist today

Share