How to Slash Data Center Energy Costs: 36 Green Ops Strategies and Real-World Cases

For IT operations staff, system stability and security are top concerns, yet the energy consumption of data centers is often overlooked. Statistics show that electricity accounts for 60‑70% of total data‑center operating costs, with air‑conditioning alone consuming about 40% of that electricity.

Green IT Must Be Actionable

Energy‑saving in data centers requires a holistic approach: site selection and building design, airflow organization and cooling systems, green power and power‑distribution, as well as IT, lighting, and cleaning practices.

DevOps "36 Strategies" – Energy‑Saving Operations Chapter

The following strategies are highlighted (selected from the full DevOps guide):

Design fully enclosed data centers to improve room sealing (no external windows, sealed doors, limit personnel access).

Use light‑colored exterior walls and insulate cooling zones to reduce thermal load; avoid glass walls.

Implement hot‑aisle/cold‑aisle separation or enclosure to prevent short‑circuiting of airflow.

Adopt energy‑efficient lighting and intelligent lighting control to reduce unnecessary illumination.

Case Study 1 – Large IT Enterprise Data Center

Key Strategies Applied: 2, 3, 8, 34 (as listed above).

Background: In 2002 the author helped build a Beijing data center that used glass walls and arranged equipment by business application. The design resulted in a PUE > 2.9.

Analysis:

Glass walls, chosen for aesthetics, acted as excellent heat conductors, raising energy use.

Grouping equipment by application created uneven heat loads and complicated cabling, leading to inefficient cooling.

Head‑to‑tail rack placement caused hot exhaust from front racks to feed directly into intake of rear racks.

Frequent visitor access leaked cool air and introduced dust, increasing energy use and security risk.

Reflection: The root cause was a lack of awareness of airflow organization and energy‑saving planning; aesthetics and safety were prioritized over efficiency.

Solutions Implemented:

Eliminate windows; construct insulated solid walls.

Redirect visitor tours to a control center instead of the server hall.

Adopt modular data‑center design with separate zones (storage, PC, small‑machine) each using tailored power and cooling.

Arrange racks head‑to‑head and tail‑to‑tail to create defined hot and cold aisles.

Combine air‑cooling and water‑cooling for large‑scale facilities.

Case Study 2 – Large Petrochemical Enterprise Data Center

Key Strategies Applied: 2, 3, 8, 34.

Background: In 2016 the author visited a newly built petrochemical data center with floor‑to‑ceiling glass windows and head‑to‑tail rack layout. The design prioritized visual appeal over thermal performance.

Solutions:

Replace glass with solid walls to reduce heat transfer.

Limit lighting to meet national illumination standards; avoid over‑lighting.

Adopt hot‑aisle/cold‑aisle separation to improve airflow.

Reflection: The facility demonstrated a lack of green‑data‑center awareness; education and mindset are the primary barriers to energy‑efficient design.

Case Study 3 – Large Internet Company Data Center (Qiandao Lake)

Key Strategies Applied: 2, 3, 8, 9, 19, 20, 26, 29, 30, 27, 28, 35.

Background: In September 2015 the Zhejiang Qiandao Lake data center began operation, targeting an annual average PUE of 1.3 using lake‑water cooling, renewable energy, and micro‑module technology.

Solutions:

Primary cooling via closed‑loop lake water, reducing electric cooling demand by over 80%.

Integrate hydro‑electric, photovoltaic, and waste‑heat recovery for power and heating.

Dynamic environment management with intelligent algorithms to adjust cooling in real time.

Develop proprietary micro‑module servers, PCIe SSDs, and integrated rack solutions that consume far less power.

Adopt large‑capacity cold‑storage tanks, ice‑storage, and lake‑water cooling.

Use low‑power IT equipment and retire high‑energy legacy hardware.

Employ multi‑energy sources (solar, wind, geothermal, tidal, biogas) to lower carbon emissions.

Implement combined cooling‑heating‑power (trigeneration) systems.

Leverage cloud computing to share resources, eliminate idle workloads, and retire “zombie” applications.

Results: The lake’s stable 17 °C deep‑water temperature enables the data center to operate 90 % of the time without external cooling, saving tens of millions of kWh annually and reducing carbon emissions by over 10 000 t of standard coal. Reported PUE averages 1.3, with a low of 1.17.

Future Outlook: China’s green data‑center initiatives are evolving from imitation to indigenous innovation, positioning the country to lead globally in sustainable data‑center design.

Source: Detailed case data referenced from CSDN article "Detailed Analysis of Alibaba Cloud Qiandao Lake Data Center".