Introduction
Updating Energy Logic Calculating the Impact Energy Logic Strategies PUE Analysis Key Takeaways Conclusion
Figure 14.The first five strategies in Energy Logic 2.0 produce a dramatic reduction in energy consumption but a small increase in PUE.
PUE
Key Takeaways from Energy Logic
These non-productive IT systems not only waste the energy they consume, they waste the energy of the power and cooling systems that support them.
Introduction
Updating Energy Logic Calculating the Impact Energy Logic Strategies PUE Analysis
Key Takeaways Conclusion Energy Logic creates a clear roadmap for driving dramatic reductions in energy consumption without jeopardizing data center
performance. As this analysis demonstrates, an organization that systematically adopted the Energy Logic 2.0 roadmap could achieve a better than 70 percent reduction in energy consumption while removing constraints to growth.
In support of Energy Logic 2.0, Emerson Network Power has created the Energy Logic 2.0 Cascading Savings Calculator at
www.EfficientDataCenters.com. Entering compute load and PUE for a particular facility allows users of the calculator to see the impact of individual Energy Logic 2.0 strategies on compute load, PUE and energy costs.
Not every organization is in a position to adopt every Energy Logic 2.0 strategy. These organizations can still benefit from the Energy Logic 2.0 analysis using these four lessons as a guide when considering data center changes.
1. Leverage the cascade effect
Because support systems account for a relatively high percentage of data center energy consumption, but contribute nothing directly to data center output, it is tempting to attack these systems first. But the load on support systems is determined by the IT load; efficiency improvements in IT systems are amplified in the support systems. In addition, a relatively high percentage of IT systems contribute almost nothing to data center output. This turns the cascade effect on its head — these non-productive IT systems not only waste the energy they consume, they waste the energy of the power and cooling systems that support them. In fact, in the base Energy Logic data center, an unutilized 200 W server actually consumes 381 W of power. That means every Watt of stranded server capacity is actually is wasting 1.91 W of energy (Figure 15).
Server Component
-2.84 W 200 W
200 W Idle Server wastes 200 W
Power Distribution
200 W
Plus 5.4 W here
UPS 205.4 W
Plus18.8 W here
Cooling 224.2 W
Plus143.6 W here
Switchgear/
Transformer 367.8 W
Plus13.4 W here
Introduction
Updating Energy Logic Calculating the Impact Energy Logic Strategies PUE Analysis
Key Takeaways Conclusion
Figure 15. An unused 200 W server actually consumes 381 W of facility power when support system energy consumption is considered.
A 200 Watt Idle Server Wastes 381.2 Watts of Power at the Utility Entrance*
* Assumes a PUE of 1.91
Excludes site and building generator, natural gas/diesel, water, fire, security, etc.
“Reverse” Cascade Effect
Energy Logic 2.0 demonstrates that huge reductions in data center energy consumption are possible using proven technologies that do not impact the data center’s ability to deliver services.
Introduction
Updating Energy Logic Calculating the Impact Energy Logic Strategies PUE Analysis
Key Takeaways Conclusion 2. Don’t compromise availability and flexibility for efficiency
Data center energy consumption has created a problem for data center-dependent organizations — and an opportunity for companies seeking to market solutions to that problem. Unfortunately, many of these “solutions” put efficiency above availability and flexibility, which is both dangerous and unnecessary. Energy Logic 2.0 demonstrates that huge reductions in data center energy consumption are possible using proven technologies that do not impact the data center’s ability to deliver services.
3. Higher density equals better efficiency
While many of today’s data center managers have spent the majority of their careers managing facilities with densities well below 5 kW per rack, the servers and support systems of today are not only capable of being deployed at much higher density, they are designed for it. While a density of 20 kW per rack may make some data center managers nervous, today’s data center can be designed to safely and effectively support that density — with room to grow.
4. Capacity is the flip side of efficiency
Despite rising costs, electricity is still relatively cheap in some areas. That has prevented some organizations from aggressively moving to optimize efficiency. However, the Energy Logic roadmap is more than a solution to rising energy costs; it is a solution to increasing demand for compute capacity. Energy Logic can eliminate the need for expensive build- outs or new facilities by removing constraints to growth as the demand for compute and storage capacity continue to grow.
Conclusion
While a density of 20 kW per rack may make some data center managers nervous, today’s data center can be designed to safely and effectively support that density—with room to grow.
Introduction
Updating Energy Logic Calculating the Impact Energy Logic Strategies PUE Analysis
Key Takeaways Conclusion The original Energy Logic brought clarity to data center energy efficiency discussions by taking a statistical approach that highlighted
the overriding importance of IT equipment in determining data center efficiency. This was an important development in the evolution of data center efficiency; however, economic conditions and a lack of real-time visibility into IT and facilities systems has limited the progress that has been made in optimizing data center performance. What gains have been made have largely been consumed by increases in capacity.
Energy Logic 2.0 illustrates the potential that still exists to optimize the data center, showing how the energy consumption of a
“typical” 5,000 square foot data center could be cut by more than 70 percent using available technologies. With effiency remaining a priority and the introduction of a new generation of management systems that provide greater visibility and control of data center systems, the time is now for the industry to begin making large strides in reducing the overall energy consumption of data centers.