A media/information company decided to consolidate their traditional servers into blade servers to conserve space, curb energy costs, and avoid construction of a new facility. They also needed a faster, more efficient way to add assets to their data center without fear of overload. To accomplish these goals, the customer knew that real-time monitoring of circuits was essential to avoid down time.
Enter real-time monitoring
Nlyte Energy Optimizer, (formerly FieldView) a part of the Nlyte suite of data center solutions, is a software solution that helps all types of data centers, especially large environments, run at peak efficiency. It provides real-time information about all the critical systems in a data center — both IT and facilities systems — in an easy-to-use format that empowers management to make great decisions based on facts, not guesses. Users get the most out of the equipment and space available, reduce energy and cooling consumption, and proactively plan for future expansion.
Nlyte Energy Optimizer’s real-time branch circuit monitoring keeps customers constantly aware of the situation in their data centers.
The software generates reports on each circuit and alarms when thresholds are approached, so changes can be made before disaster strikes. This allows consolidation without overload.
The solution was chosen for its ability to monitor circuit usage in real time. Common practice in data centers that don’t have real- time monitoring is to leave a large margin for error, known as over-provisioning, in case of spikes in demand. The customer determined that roughly 80% of the servers in their three data centers were only utilizing 10% to 15% of their CPU capacity.
As the computing load was transitioned to the more efficient blade servers, Nlyte Energy Optimizer’s real-time monitoring allowed the customer to close this “safety gap” in compete confidence that thresholds would not be reached or exceeded.
As traditional servers transition onto blade servers an average consolidation ratio of 9 to 1 is achieved. At an average workload, a traditional server consumes 500 Watts of power vs. about 2,500 Watts per blade server. This means a 45% reduction in direct energy costs for the same amount of processing while increasing fault tolerance and performance through load balancing. With a typical PUE of 2.0, costs associated with cooling and other infrastructure support are also roughly halved.
One thousand two hundred forty three traditional servers were consolidated onto 137 blade servers. The resultant reduction in energy costs was over $500,000. Based on drawing power from a coal-fired base-load power plant, the reduction in carbon emissions from this project equates to 2,841 tons. The consolidation would not have been possible without Nlyte Energy Optimizer to constantly monitor the load.
In addition to this conservation of both space and power, introducing new assets became much easier as well. Instead of the time-consuming tasks of manually checking each cabinet load, then having to do calculations, which can take a week, only to produce outdated data, Nlyte Energy Optimizer software constantly and automatically monitors load so availability can be discovered in a matter of minutes.
The customer plans further blade server consolidation. With the Nlyte software solution, these changes will continue to run smoothly, helping make the data centers more efficient.