WaterWise: Co-optimizing Carbon- and Water-Footprint Toward Environmentally Sustainable Cloud Computing

TL;DR

WaterWise tackles the environmental challenge of cloud computing by co-optimizing carbon and water footprints across geographically distributed data centers. It introduces a MILP-based scheduler that leverages delay tolerance and inter-region transfer to exploit temporal and spatial variations in carbon and water intensity, incorporating a water-scarcity aware framework with offsite and onsite water components. The approach demonstrates substantial reductions in both carbon (~21%) and water (~14%) footprints over a baseline and remains effective under varying delay tolerances, regional data, and workloads, with low decision overhead. This work provides an interpretable, open-source framework for sustainable cloud management and highlights the essential coupling of energy-water trade-offs in real-world deployment.

Abstract

The carbon and water footprint of large-scale computing systems poses serious environmental sustainability risks. In this study, we discover that, unfortunately, carbon and water sustainability are at odds with each other - and, optimizing one alone hurts the other. Toward that goal, we introduce, WaterWise, a novel job scheduler for parallel workloads that intelligently co-optimizes carbon and water footprint to improve the sustainability of geographically distributed data centers.

Figures (13)

• Figure 1: Carbon intensity and water requirements (Energy Water Intensity Factor (EWIF)) for energy generation for different types of energy sources. Carbon-friendly energy sources can have higher water needs for energy generation (EWIF), contributing to a higher offsite water footprint.
• Figure 2: Carbon Intensity, EWIF, WUE, WSF varies across different geographical regions (average values for the year 2023, region labels are sorted according to the carbon intensity). Carbon intensity and water intensity show temporal variations.
• Figure 3: Quantifying optimal solution benefits, opportunity scope due to delay tolerance, and impact on job distribution across regions.
• Figure 4: Design overview of WaterWise.
• Figure 5: WaterWise provides significant carbon and water footprint savings compared to the baseline. Higher delay tolerance can further improve carbon and water footprint saving.