Droughts, Heat Waves, and High Data Center Cooling Costs

Jeff Klaus<br/>IntelJeff Klaus

Jeff Klaus is General Manager of Data Center Solutions at Intel Corporation.

Summer is not all fun. In the data center, IT and facilities teams are happy to see an end to summer and the extra strain it puts on the air handlers and cooling systems. Finance teams similarly celebrate an end to the higher utility bills.

As autumn approaches, it is a good time to evaluate the effectiveness of the in-place cooling solutions, and consider changes that can cut costs and help everyone keep their cool next year.

The $$$$ of Climate Control in the Data Center

Lowering cooling costs should never start with blindly raising the set point on the thermostat in server rooms. Yes, bumping it up a single degree can translate into a noticeable reduction of cooling costs. However, temperature in the data center is never uniform. It can vary drastically throughout a room, a row, or even within a single rack.

At a minimum, undetected hot spots must be identified and addressed before making any set point changes, which calls for a tool for monitoring temperature patterns. Fortunately, data center equipment providers have built in the intelligence that supports fine-grained thermal and power monitoring. Rather than relying solely on the return-air temperature at the cooling systems, site managers can benefit from temperature and power data provided by individual servers, blades, power distribution units, air handlers, and other intelligent devices in the data center.

Continual Monitoring – Automatically

An energy management solution can automate the ongoing collection and aggregation of these data points. IT and facilities teams can then view and study real-time thermal conditions as well as long-term patterns.

The best-in-class energy management solutions feed a steady stream of this data to a console, with various display options including at-a-glance thermal maps. Once hot spots are exposed, adjustments can be made to cooling systems or rack densities for immediate improvements in temperature consistency and cooling efficiency.

Over time, monitoring and analyzing logged temperature data can provide insights about the correlations between cooling costs and outside weather and environmental conditions.

IT can also monitor the ongoing efficiency of cooling systems to more optimally schedule preventative maintenance. In data centers that employ water-based cooling systems, these oversights of cooling efficiency can directly lower water consumption and help comply with any restrictions imposed during droughts.

Mitigating Cooling System Failures

The most energy efficient data centers constantly check for under- or over-cooled spots since these are signs of flaws in the cooling plan. A sudden increase in temperature can also alert the data center team to a failing cooling system.

The cooling plan should include steps that will be taken to mitigate any such cooling system failures. These same steps can apply during periods of power outages or restrictions, or any time that the temperature in the data center starts to increase.

The mitigation plan might include bringing in outside air (via economizers, or other ventilation), setting up fans (if power is available), or taking non-essential systems and services offline to reduce demand. Thermal monitoring capabilities and analysis of temperature correlations make it possible to define a mitigation plan that will optimize results under any conditions.

The combination of ongoing monitoring and a detailed failure plan puts IT and facilities teams in a position of readiness, and also promotes proactive maintenance practices that minimize server failures due to hot spots and cooling system failures.

Go Ahead – Turn Up the Temperature

With an understanding of the temperature patterns and correlations, IT can now confidently raise the set point in the data center. While most data centers still operate at 70 degrees Fahrenheit or lower, growing numbers of sites are pushing the data center temperatures up to 80, 82, or even higher set points. With each degree of increase, power bills reflect a two-percent increase in savings, which adds up significantly year after year.

There is no magic temperature setting rule, and every data center team must still consider the local climate conditions, budget, and equipment requirements. Today’s servers can definitely handle higher operating temperatures, but some storage systems and legacy tape backup systems are more sensitive to temperature and humidity. Adjusting room layouts and consolidating sensitive equipment in cool zones can still allow higher ambient temperatures and, therefore, cooling reductions in major portions of any data center.

The same monitoring solutions that automatically collect and aggregate temperature data can support threshold definitions and automatic alerts and actions to protect equipment and maintain conditions that maximize the lifespan and reliability of critical equipment.

Be Cool

By adopting the temperature monitoring practices of the world’s largest and most energy efficient data centers, even a small data center can achieve cost savings and improved uptimes. Before next summer has IT and facilities sweating about cooling solutions, consider a free trial or pilot deployment of a real-time temperature monitoring solution. Better yet, look at a solution that combines temperature and energy monitoring. Very cool.

Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.

Add Your Comments

  • (will not be published)

One Comment

  1. You should also integrate the water quality and water efficiency of any data center cooling system. It is egregiously overlooked. Water quality directly impacts the overall efficiency and performance of your cooling system; if you're not continuously monitoring the water quality 7x24, you could be headed for downtime or worse: a complete critical failure.