System-Level Impacts of Flexible Data Center Load Scheduling on Cost, Emissions, and Transmission Congestion
New study shows BE jobs shifting to low-LMP hours reduces grid stress.
A new study from Akibul Hasan Mazumder and Yuanrui Sang, published on arXiv, explores how flexible scheduling of data center workloads can benefit the power grid. Using the ACTIVSg2000 2000-bus test system, the researchers modeled the impact of shifting best-effort (BE) jobs—tasks like batch processing or data backups—to times when electricity prices are lower. They found that BE loads naturally migrate to periods with lower locational marginal prices (LMPs), which often align with high renewable energy output, such as sunny or windy days.
This scheduling flexibility leads to significant system-level benefits: reduced operating costs for data centers, lower greenhouse gas and toxic emissions, and decreased transmission congestion. Crucially, latency-critical (LC) workloads—like real-time AI inference or video streaming—remain unaffected, preserving quality of service (QoS). The study demonstrates that even modest flexibility in data center operations can support more efficient and sustainable grid operation, offering a practical pathway for data centers to contribute to decarbonization without compromising performance.
- Flexible scheduling of BE jobs shifts load to low-LMP periods, often coinciding with high renewable generation.
- Results show reductions in operating costs, emissions (both greenhouse gas and toxic), and transmission congestion.
- Latency-critical workloads remain unaffected, preserving quality of service (QoS).
Why It Matters
Data centers can reduce grid costs and emissions by scheduling non-urgent tasks during renewable-rich hours.