By Anjani Kumar, Country Manager, India & SAARC, Raritan- A brand of Legrand
Given that data centers run so many mission critical applications, it is important that businesses have watertight processes in place to closely monitor their operations at all times. Yet, this doesn’t always happen in practice. In a bid to cut costs, businesses sometimes don’t give data center monitoring the priority that it deserves, thereby creating unwarranted risks that threaten business continuity. Not only this, they run the risk of harming their entire infrastructure. Data center heating can be a major issue especially in a country like India where ambient temperatures are already high. Adhering to certain best practices for data center monitoring can help mitigate these risks and ensure ideal environment for the data centers.
Often, data center managers use a checkbox approach to monitoring. For example, temperature sensors may have been installed within the room that houses the data center but there would have been no thought given to if the number of sensors are adequate or whether these sensors are accurately positioned convey an accurate picture. Most experts recommend that each rack should have least six temperature sensors. Having too few sensors which only monitor the overall temperature of the room is a recipe for disaster. For instance, some portions of the data center may be more at risk than others. Unless the sensors are able to give an accurate picture, they cannot be relied upon. Along with the number of sensors, their positioning matters too. There should be sensors at the top, middle, and bottom at the front and back of each rack. This way, data center managers can detect any heating issues right to the rack level.
Having only overall room monitoring does not give an accurate picture because alerts are raised only at the point when the air conditioning units in the room are no longer able to compensate for the additional heat. The situation might already be unmanageable at this stage. The business might have to contend with downtime or rack damage as it struggles to resolve the issue at this stage.
Even if a business is unable to adhere to the six sensors per rack guideline for cost reasons, at least 3 sensors per rack should be considered a bare minimum to maintain optimal safety. These can be placed at the bottom front of the rack, the top front of the rack, and the top back of the rack. It is crucial to measure the top back of the rack since this point generally tends to heat up quickly and needs careful monitoring. The sensor at the top front of the rack can be a good measure of the cooling system since it can indicate whether the cold air is reaching the top of the rack. Similarly, bottom front sensor could be used in conjunction with airflow monitoring to ensure that cold air is reaching the rack.
The recommended range for intake temperatures is between 18°-27°C and outtake temperatures should be less than 20°C as compared to the respective intake temperature. Accurate monitoring of the intake and outtake temperature plays an important role to ensure maximum uptime. At the same time, it also increases lifetime expectancy of data center equipment. While high end systems come equipped with auto shutdown capabilities not all models come with such safeguards. Overheating can impact equipment life expectancy and put system availability is at a higher risk.Of course, business continuity suffers too. By embracing best practices and using the right tools, data center managers can get a much better handle on monitoring of their IT infrastructure.
Also, when it comes to effective monitoring, using the right tools facilitate effective environmental monitoring besides delivering optimal insights in a cost effective manner. Thus greatly bringing down the risk of overheating.
If you have an interesting article / experience / case study to share, please get in touch with us at [email protected]