Reduce Datacenter Costs by Monitoring Energy Consumption
When it comes to datacenter maintenance, enterprise organizations face a constant challenge – how to increase efficiency, while also driving down operating expenses?
Aside from specific individual business challenges, datacenters also get caught up in the never-ending circle of compute capability, power consumption and cooling.
In response to these mounting pressures, many enterprises are seeking solutions to increase efficiency while cutting costs in their datacenters.
IDC surveyed 404 datacenter managers and operators across the United States and recently published their findings in "Devouring Power in the Datacenter: Results from IDC's Enterprise Data Survey."
According to the survey, the average datacenter budget for enterprises is $1.2 million. The highest percentage of datacenter budgets – 24 percent – is spent on power and cooling, with an average spend of $300,000. More and more, enterprises are figuring out ways to leverage power and cooling to drive down costs.
Most frequently, datacenters are turning to airflow management solutions such as blanking panels and venting tiles. Other cost-cutting solutions include liquid cooling and enabling variable speed fan drives. Some datacenters have even increased temperatures from 70-75 degrees to 78 degrees to combat costs.
One solution that might yield a high cost reduction for datacenters is the implementation of intelligent Power Distribution Units (PDUs), which deliver power to IT equipment and provide energy consumption data.
Intelligent PDUs provide granular energy and environmental monitoring. They also prevent an over-provisioning of power, which could restrict growth in the data center
Discovering Excess Server Capacity
A recent study from Stanford University research fellow Jonathan Koomey and Anthesis Group partner Jon Taylor further points to the need for more efficiency in datacenters. According to their research, server utilization in business and enterprise datacenters rarely exceeds six percent, meaning they deliver no more than six percent of their maximum computing output over the course of the year.
The Uptime Institute, which was cited in their research, reports that an estimated 30 percent of physical servers in the United States are comatose, meaning they haven’t done anything for six months.
Many of these same servers – about 3.6 million in the U.S. and 10 million worldwide – are draining lots of energy, despite being inactive.
In their research, Koomey and Taylor said managing the datacenter’s resources by CPU memory usage alone won’t ensure that a particular server is comatose.
"Data centers often have excess capacity because identifying unused or over-provisioned hardware with certainty is difficult with conventional tools," Koomey and Taylor said.
"Data center operators also need to look at upstream traffic or user access information per server from central IT management, virtualization and workload distribution systems. This approach identifies IT resources not doing any useful work so they can be decommissioned without adding risk to the business."
Creating Value with Performance Monitoring
When thinking about increasing efficiency and reducing costs in your datacenter, consider a digital infrastructure monitoring platform that has the ability to correlate energy consumption data with IPMI data and other performance metrics. This type of insight provides a better understanding of:
- Compute power consumption
- Opportunities for server virtualization and server consolidation to reduce power consumption and cooling
- Potential operational efficiencies through better utilization of available server resources
Automatic baselining of energy data can help enterprises understand normal power consumption and ferret-out unusual conditions, even across hyperscale datacenters
For advice on virtualization in the datacenter, watch our video, “Identifying Network Technology and Server Candidates For Virtualization.”