What is it about?
Databases are everywhere—from mobile apps to scientific research—and they often run on large servers that consume a lot of electricity. Most databases are designed to process queries as fast as possible, using full power throughout. But what if they could plan ahead and reduce power use when possible? In this study, we designed a new method that allows a database system to preview each query, identify which hardware parts are actually needed, and power down the rest. It’s like turning off the lights in rooms you’re not using. We implemented this approach on PostgreSQL, a popular open-source database, and tested it using standard benchmarks. Our results show that this proactive energy management can cut electricity use by up to 29% while slowing queries by only about 8%. This makes it possible to reduce the environmental footprint of data centers—without needing new hardware or major trade-offs in performance.
Featured Image
Photo by Kirill Sh on Unsplash
Why is it important?
Data centers already use about 1–2% of the world’s electricity, and every database query—from big data analysis to everyday apps—adds to that demand. Yet most database systems are still built to focus almost entirely on speed, with little regard for how much power they use. Our research makes energy a key part of database software design, not just an afterthought. By letting the system look ahead at how a query will run, it can selectively power down hardware components that aren’t needed—saving electricity without major performance loss. In our tests, this proactive approach reduced energy use by up to 29%, with only an 8% slowdown. Because our method is entirely software-based, it can be deployed through regular updates—no new hardware required. At scale, this could mean lower energy costs and smaller carbon footprints for millions of servers worldwide, making it a timely and practical step toward more sustainable computing.
Perspectives
Energy efficiency in database systems drew attention in the late 2000s, but the momentum faded before the field could fully mature. That always seemed like a missed opportunity to me—especially given the growing energy demands of data centers and the central role databases play in so many systems. This project began with a simple idea: what if the database executor could think about energy, not just performance? That question led me down a path of reimagining how a query engine could act more deliberately—planning ahead and making power-aware decisions on its own. I see this work as just a small step toward a broader vision: making energy a first-class concern in system design. There's still a long way to go, and I hope to continue developing this into a more comprehensive and practical approach to sustainable data processing.
Project research associate Yuto Hayamizu
The University of Tokyo
Read the Original
This page is a summary of: Proactive Energy Management in Database Systems, ACM SIGEnergy Energy Informatics Review, December 2024, ACM (Association for Computing Machinery),
DOI: 10.1145/3727200.3727223.
You can read the full text:
Contributors
The following have contributed to this page
