Energy Efficiency in Data Centers
LSI Research
Energy efficiency in data centers is a major challenge due to its economic and environmental impact. At LSI we develop proactive strategies that coordinate applications, compute and cooling resources to reduce total consumption.
2024-2025 snapshot
Data center energy demand keeps accelerating and is now a visible share of global electricity.
Projection and efficiency
These projections highlight how urgent efficiency innovation is to sustain growth without straining grids, costs, or climate goals.
GreenLSI
The GreenLSI team is part of the Integrated Systems Laboratory (LSI) in the Department of Electronic Engineering at Universidad Politecnica de Madrid. Our research focuses on Energy Optimization in Data Centers.
We developed a global strategy that integrates multiple data sources to coordinate decisions and reduce total energy use. The system combines knowledge about application energy demand with compute and cooling characteristics to apply proactive optimization techniques.
Decision framework
The system uses power models trained with real data, then evaluates optimization results with a decision-making module that proposes actions to be executed back in the data center.
Multi-layer optimization
Our framework gathers information across the data center to generate knowledge that enables optimizations at multiple abstraction levels: resource manager, architecture, compiler, and more. This holistic view yields coordinated, measurable decisions.
2-Phase Immersion Cooling
Immersion cooling technology applied to reduce energy consumption in data centers.