How to Implement Energy-Efficient Data Centers
Release time:
2017-10-16
Electricity Consumption—An Effective Metric for Measuring Data Center Operational Efficiency. Throughout the entire lifecycle of a data center (8 to 10 years), electricity costs can exceed both the costs of IT equipment and system acquisition costs. By taking appropriate considerations during the construction phase, it’s possible to reduce electricity costs by 20% to 50%. Moreover, if a more systematic planning and design approach is adopted, electricity consumption can be reduced by as much as 90%! Today, people typically use data center efficiency to gauge the operational quality of a data center; however, the concept of data center efficiency itself is rather vague. If we shift the measurement standard to electricity consumption instead, the assessment criteria will become much clearer and more precise.
Electricity Consumption—An Effective Metric for Measuring Data Center Operational Efficiency
Over the entire lifecycle of a data center—typically 8 to 10 years—electricity costs can exceed both the cost of IT equipment and the cost of system acquisition. By taking appropriate considerations during the construction phase, it’s possible to reduce electricity costs by 20% to 50%. Moreover, if a more systematic approach to planning and design is adopted, electricity consumption can be reduced by as much as 90%!
Today, people typically use data center efficiency as a metric to gauge the operational quality of data centers. However, the concept of data center efficiency is actually quite vague. If we shift the measurement standard to electricity consumption instead, the evaluation criteria become much clearer. In China, the cost per kilowatt-hour usually ranges from 0.68 to 0.70 yuan; thus, 10 kilowatts would cost roughly between 6,000 and 7,000 yuan. For a typical data center operating over a 10-year period, the electricity costs associated with each kilowatt of load would amount to between 60,000 and 70,000 yuan. Therefore, where exactly this electricity is being consumed is an important factor that needs to be carefully considered in the construction, design, and operational management of data centers.
According to statistics, only about half—or less than half—of the energy consumed by data centers is actually used by IT equipment; the rest is consumed by infrastructure. There are two main approaches to help reduce energy consumption: one involves temporary measures, and the other focuses on structural measures. The former includes reducing server loads, decreasing the number of servers, or implementing energy management systems for servers. However, these temporary measures address only the symptoms rather than the root causes of the problem—they are merely stopgap solutions. In contrast, structural measures can maximize energy savings and permanently reduce power consumption. Estimates suggest that permanent measures can save at least twice as much electricity as temporary measures. Reducing the power consumption of IT systems is a key strategy for implementing the first approach. Such measures include retiring outdated and unused equipment, adopting efficient management practices, switching to high-efficiency systems, and leveraging virtualization and standardized designs. The second approach primarily relies on designing highly efficient data center infrastructure systems and using high-efficiency UPSs and cooling equipment. When assessing data center energy efficiency, system design plays a more critical role than the selection of power supplies and cooling equipment. The efficiency differences among data centers built with different designs can be enormous. Depending on the workload, properly configuring the physical infrastructure of a data center has the greatest impact on power consumption. According to statistics, proper configuration can potentially reduce electricity costs by as much as 50%.
Best Practices for Data Center Energy Efficiency: According to APC, there are at least ten ways to save energy in data centers:
1. By rationally configuring the physical infrastructure of data centers, energy consumption can be reduced by 10% to 30%.
2. Adopting server virtualization can reduce energy consumption by 10% to 40%.
3. Adopting efficient cooling modes and architectures—such as row-level and cabinet-level cooling, as well as hot aisle containment systems—can save energy consumption by 7% to 15%.
4. By using the energy-saving mode of air conditioners, energy consumption can be reduced by 4% to 15%.
5. By adopting a more efficient floor layout, energy consumption can be reduced by 5% to 12%.
6. Adopting more efficient power supply equipment can save energy consumption by 4% to 10%.
7. Coordinating the operation of air conditioners can save energy consumption by 0–10%.
8. Properly arranging the air outlets can save energy consumption by 1% to 6%.
9. Installing high-efficiency lighting equipment can save energy consumption by 1% to 3%.
10. Installing blind flanges can save energy consumption by 1-2%.
In short, building an efficient data center requires efforts in the following areas: conducting preliminary planning, adopting standardized processes and procedures, comprehensively considering business-level impacts and identifying IT requirements, implementing effective layout and design, specifying equipment parameters, and finally, ensuring smooth operational implementation. In the long run, the investment made in building an efficient data center will far outweigh its benefits.
Keywords:
Next page
Previous page:
Next page:
Recommended News
2018-08-28
Provide consumers with more solutions.
Power by : 300.cn Beijing | CertificateQuery | File information inquiry | Anti-counterfeiting query