Emerson Network Power Considers How Far Organizations Should Go to Improve Data Center Efficiency

Accordingly, the experts at Emerson Network Power now are frequently asked to help organizations determine how to take advantage of these improvements, especially in data centers that have run or are running out of capacity. What’s the better course of action? Should they retrofit the existing data center or replace it with a new one?

When to Retrofit

A retrofit is a good option when efficiency and stranded capacity are the major issues. “Retrofitting has advantages over building a new data center, cost being at the top of the list,” said Peter Panfil, vice president of global power, Emerson Network Power. “Today, a new data center could cost in the range of $17M per megawatt, with ways to push this down to $12M per megawatt (for example, through UPS utilization improvements). A retrofit will run 25 percent to 40 percent of the new-build cost and in many cases can deliver the efficiency and capacity needed to extend the life of the data center.”

Newer cooling, power and IT technologies that improve energy efficiency also help recover stranded capacity. Recovered capacity then can be allocated where needed.

Cold-aisle containment, intelligent controls that monitor conditions at the server and economization all lead to better efficiency by improving temperature and airflow management.

Additionally, implementing component technologies that optimize the cooling system to work more efficiently at partial loads is another good way to recover capacity without starting from scratch. Variable speed drives improve efficiency significantly over traditional fixed-speed fans, and electronically commutated (EC) plug fans are even more efficient. Both variable speed drives and EC plug fans can be installed on existing cooling units and their operation optimized by networking the cooling units through intelligent controls. Intelligent controls ensure the units work together as a system to optimize performance and efficiency.

“Component upgrades such as these are much less intrusive than replacing entire cooling systems, and their payback is typically about two years,” said Steve Madara, vice president and general manager, Emerson Network Power’s Liebert Precision Cooling business. “When you consider adding new components to upgrade existing precision cooling equipment, however, keep in mind that it’s like deciding when to trade in your car for a new model. There’s a certain point where you can’t get enough mileage from retrofitted equipment to make it worth the investment.”

Upgrades that make the power system more efficient include double-conversion UPS technology that eliminates conversion process losses, intelligent pairing of UPS modules and high-efficiency transformers. Additionally, adding a data center infrastructure management solution, implementing eco-mode on the UPS and optimizing the power path from the UPS to the servers all contribute to energy savings.

Data center managers also can employ Emerson Network Power’s Energy Logic to improve energy efficiency. Energy Logic leverages the cascade effect that occurs when lower energy consumption at the component and device level is magnified by reducing demand on support systems. It’s possible to achieve high overall energy savings from the cascade effect. Saving energy extends the useful life of the data center by freeing space in the data center that can be used for IT expansion.

When to Replace

While many options for retrofitting the data center exist, it is not always the practical choice. “The downside of a major retrofit is that you are doing it in a live data center,” Panfil said. “If you are maxed out on space, it’s difficult to upgrade for efficiency without jeopardizing availability.” In this case, upgrading unit by unit may protect against downtime, but organizations should consider building a new data center when:

1. The existing space is inadequate to support new, more efficient equipment.
2. The data center is out of space but can be put into service as a disaster recovery site.
3. The organization has multiple data centers and wants to improve utilization. The data centers can be consolidated over time into a new data center that has all the latest efficiencies.

Alternatives

If a retrofit is not an option and a complete new build is not feasible, organizations still have alternatives. When high-availability organizations run out of options, one alternative is going to a collocation facility for overflow and demand management of their IT requirements. They can rely on that strategy for applications with lower criticality, such as email, and keep applications supporting their core competency in-house. Building additional space to handle more efficient technologies and moving high-density applications into it is another option.

Alternatively, when gaining efficiency and capacity are the goals, deploying a containerized/modular data center to house high-density applications might be a good option. Containers can be equipped with the latest high-efficiency technologies, and future capacity can be gained by adding more containers – at less cost and risk than retrofitting or building new.