3 ways to eliminate data center hotspots
Tuesday, May 21st 2013
Optimizing environmental monitoring for data centers is no easy task. There are many factors to consider, and an oversight in one or more areas can have serious consequences. Should a calamity strike, data center operations could decelerate or even shut down, costing enterprises valuable time and resources. Hotspots are one risk factor to note, as the heat that insufficient cooling on a high-power area can generate can wreak havoc on equipment. Hotspots are much easier to prevent than they are to fix. Taking a proactive approach to data center monitoring can pay significant dividends, as a data center with zero downtime is one of the foundations of ideal organizational execution. Here are three ways to minimize the dangers of hotspots.
1. Location, location, location
One important point to remember about hotspots is that they are acute anomalies, not tiny deviations from the norm. A disregard for the fluctuation potential of hotspots is risky, cautioned Mo Sheikh, marketing manager at environmental monitoring solutions provider ITWatchDogs. While the computer room's ambient room temperature measurement should be taken into account, hotspots can be much hotter in different parts of the room and inside server cabinets. It is important that temperature monitoring includes measurements from all around the data center. Accurate measurements can help promote long-term data center health as well, according to Network World. Thermal and power data aggregated over time could be applied to a cartographic picture of the data center, in which hotspots and power users are singled out for optimal resource allocation and preventative device cooling.
2. Sensors and sensibility
Environmental monitoring equipment placement is crucial, as poorly located sensors can fail to give an accurate reading of the room's hotspots and even miss them altogether. Sheikh advocated a comprehensive approach to sensor placement.
"It is critical to position sensors in various locations - usually top, middle, and bottom - of a server rack and around any heavy duty machinery," he stated.
This way, the readings that temperature monitoring systems produce can be focused and specific, allowing for changes in airflow that might affect upper server monitoring levels and cooling agents that could cause the server parts located closer to the ground to fluctuate.
3. Take action
If a large-scale remodeling or reconfiguring of servers is not an optimal option, then data center operators have several other ways to combat hotspots without having to change everything. One is to leverage 'point cooling,' which involves installing supplemental cooling devices next to hotspots. Monitoring IT hardware should continue to be a top priority, and comprehensive monitoring devices like the ones ITWatchDogs offers will help facilities managers stay on top of hotspots.