Skip to content

Export Citation

        
          
@article{fieni-2024,
TITLE = {{PowerAPI: A Python framework for building software-defined power meters}},
AUTHOR = {Fieni, Guillaume and Acero, Daniel Romero and Rust, Pierre and Rouvoy, Romain},
URL = {https://hal.science/hal-04601379},
JOURNAL = {{Journal of Open Source Software}},
PUBLISHER = {{Open Journals}},
VOLUME = {9},
NUMBER = {98},
PAGES = {6670},
YEAR = {2024},
MONTH = Jun,
DOI = {10.21105/joss.06670},
KEYWORDS = {Power ; Energy ; Toolkit ; Measurement ; Framework},
PDF = {https://hal.science/hal-04601379/file/10.21105.joss.06670-3.pdf},
HAL_ID = {hal-04601379},
HAL_VERSION = {v1},
}

About Us

History

PowerAPI was initiated in 2011 by the ADAM project team at the Inria Center of the University of Lille. The primary motivation was to create a tool that could estimate the energy consumption of software processes without relying on external hardware like wattmeters. This need arose from the limitations of existing tools, which often lacked configurability and required hardware-based measurements.

The initial version of PowerAPI was developed in Scala, utilizing the Akka library for its modular and asynchronous event-driven architecture. This design allowed users to monitor energy consumption at the process level by analyzing hardware resource utilization such as CPU, memory, disk, and network usage. The modularity of PowerAPI enabled users to customize their monitoring setups by adding or removing components as needed.

Between 2015 and 2020, PowerAPI saw significant growth in both functionality and adoption. PowerAPI's architecture was refined to support more diverse monitoring scenarios, including virtual machines and containerized environments.

Since 2021, recognizing the need for broader accessibility and integration with modern software ecosystems, the PowerAPI team initiated a transition from Scala to Python. This shift aimed to leverage Python's widespread use in data science and machine learning communities, facilitating easier adoption and extension of the toolkit.

The Python-based version of PowerAPI introduced several enhancements:

  • Improved Extensibility: The new architecture allowed for easier integration of custom sensors and estimation models, accommodating a wider range of monitoring scenarios.

  • Real-Time Monitoring: PowerAPI continued to provide real-time energy consumption estimates, now with improved accuracy and support for additional data sources.

  • Integration with Visualization Tools: The toolkit offered compatibility with visualization platforms like Grafana and Prometheus, enabling users to create comprehensive dashboards for energy monitoring.

In June 2022 a Consortium was created for supporting this development and maintenance with 4 founding members: Inria, Orange Labs, Ovh Cloud and Davidson Consulting. The continued evolution of PowerAPI reflects its commitment to advancing energy-aware software development practices.

Consortium members

Inria
Since June 2022
Orange Labs
Since June 2022
OVH Cloud
Since June 2022
La Poste Groupe
Since October 2023

Past Consortium members

Davidson Consulting
From June 2022 to December 2024

How to join the Consortium

If you are interested in join the Consortium please send us an email.