Collaborative Optimization of Data Center Computing Power Based on Flexible Business Load Scheduling

Through the CCF‑Alibaba Innovative Research (AIR) Youth Fund, a joint industry‑academia platform between Alibaba Cloud and North China Electric Power University completed the project “Collaborative Optimization of Data Center Computing Power Based on Flexible Business Load Scheduling.” The project proposed a computing‑task scheduling method for low‑carbon operation of data centers, validating that renewable‑energy‑driven coordination between data centers and power systems can achieve up to 37% reduction in energy cost and 14% reduction in carbon‑related cost.

The CCF‑AIR Youth Fund, created by CCF and Alibaba, aims to provide young researchers with real‑world scenarios and resources, fostering university‑industry cooperation and technology transfer.

Started in 2021, the research focused on three aspects: mapping the relationship between business load and data‑center electricity characteristics, assessing the potential of energy‑aware scheduling, and conducting simulation and pilot verification of renewable‑energy‑driven coordination between data centers and power systems. Key outcomes include revealing dynamic mapping rules between cloud‑computing workloads and power consumption, establishing an assessment framework for scheduling potential, and proposing a low‑carbon economic operation strategy to address large‑scale heterogeneous online task scheduling under tightening energy and carbon constraints.

A pilot experiment between Zhangbei and Nantong data centers demonstrated the practical impact: under power‑system peak‑shaving signals, Alibaba Cloud’s scheduling system migrated search and recommendation workloads from the Nantong center to the renewable‑powered Zhangbei center, reducing Nantong’s load by about 100 kW and shifting roughly 150 kWh of electricity, increasing renewable consumption in the North China grid and cutting CO₂ emissions by 120 kg. This marked the first cross‑regional “computing‑power” coordinated scheduling in China, received coverage from Xinhua and other media, and contributed to Alibaba Cloud’s recognition as a national carbon‑neutral case.

Looking ahead, Alibaba Cloud will deepen these research results, expand their application across the full lifecycle of data‑center operations, build quantitative models for joint information‑energy scheduling, and further lower lifecycle energy assessment constraints, continuing to deliver stable, secure, efficient, and economical green digital capabilities to society.