Your browser doesn't support javascript. This means that the content or functionality of our website will be limited or unavailable. If you need more information about Vinnova, please contact us.

Scaling and Shifting CPU and GPU use to Reduce Energy Consumption and Peaks for Data Centers

Reference number
Coordinator Elastisys AB
Funding from Vinnova SEK 500 000
Project duration November 2022 - August 2023
Status Ongoing
Venture energy efficiency of society

Purpose and goal

We aim to reduce the energy consumption and power peaks of applications running on data centers (including both CPU and GPU resources), by dynamic scaling of resource allocations to actual needs and by shifting usage from electricity load peak hours to times of less utilization, typically during nighttime. The project will provide APIs, algorithms, and a power management framework. The project outcomes are delivered as both open specifications and open-source reference implementations.

Expected results and effects

** Denna text är maskinöversatt ** The Swedish Energy Agency estimates that Swedish data centers´ energy consumption will double from today´s 3 TWh as early as 2025. The project´s results can help shift load peaks, but also reduce total energy consumption, as studies show that data centers can save up to 50% of total by turning off non use hardware. Early results of this will be demonstrated during the duration of the project in the project partner Elastx´s data center.

Planned approach and implementation

This project, a collaboration between leading Swedish cloud provider ElastX and container platform experts Elastisys will implement CPU/GPU scaling and shifting mechanisms at infrastructure layer (ElastX) and energy-based scaling and shifting policies for container-based workloads (Elastisys). These result will be integrated and demonstrated through a representative machine learning application.

The project description has been provided by the project members themselves and the text has not been looked at by our editors.

Last updated 28 November 2022

Reference number 2022-03265

Page statistics