Reproduce “Harnessing Flexible Spatial and Temporal Data Center Workloads for Grid Regulation Services” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yingrui Fan, Junbo Zhao
Researchers at Dartmouth College developed a unified framework that co-optimizes geographically distributed data center workload scheduling with grid frequency regulation commitments. This framework demonstrably lowers overall grid operating costs while increasing data center regulation capacity and profit through robust management of real-time uncertainties.
Key insights
- Joint spatial and temporal workload shifting across data centers effectively flattens the aggregate load curve and reduces peak demand, contributing to grid stability and economic efficiency.
- The ability to dynamically redirect workloads is sensitive to latency constraints; relaxing these allows for greater economic benefits through more extensive geographical shifting.
- Robustly modeling non-Gaussian regulation signals with methods like the Gaussian envelope and managing cumulative energy deviations with VaR queue constraints is essential for reliable and sustainable data center participation in grid regulation markets.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.