russelltrow
commented
1 month ago Hi @jbellloni thank you so much for raising your project, please don't forget to register your project: https://hack.greensoftware.foundation/register/
You must register your project before you can submit your solution for judging.
jbellloni
Prize category
Best Plugin
Overview
A plugin that allows you, to apply a PUE factor. The PUE value is based on either of these inputs:
Questions to be answered
-
Have you got a project team yet?
I am working on this project alone
Project team
No response
Terms of Participation
Project Submission
Summary
PUE-Plugin applies the PUE as a factor to cpu/energy (given in the input or from teads-curve) and outputs the used pue and pue_energy. The PUE default is 1.58 (global average of large data-centers), pue can be given in config or global-config. cloud/vendor and cloud/region are optional parameters. The PUE is then imported from a csv-file.
Problems
PUE can be quite a large factor when it comes to energy consumption. Implementing this plugin can be used to calculate more realistic carbon values. Also this plugin could be used to compare different data centers and model, what effect a cloud-migration would have.
Note: There is still very little information available regaring PUE-values!
Application
described above
Prize category
Best Plugin
Judging criteria
Overall impact
better estimation of energy usage and consequently carbon demonstrates how narrow the data situation is with the major cloud providers - perhaps it will increase the pressure a little to release better data
Opportunity
integrates another factor that has not been considered so far, gives the possibility to compare coud-providers on the PUE value, otherwise does not increase the applicability of the IF
Modular
the plugin is not as modular as I would have liked... most functions are modular, still there is to much logic in the execute()
Video
https://www.youtube.com/watch?v=ivmcG_MRQMo
Artifacts
https://github.com/jbellloni/pue-plugin
Usage
sample manifest: https://github.com/jbellloni/pue-plugin/blob/main/inputs/test-manifest.yml
make sure to include cpu/energy (e.g. via the teads-curve plugin)
Process
try-and-error under stress
Inspiration
PUE was mentioned in one of the introductory videos before the hack and it's something I have been dealing with before.
Challenges
There is very little information released regarding the PUE value. GCP is the only big cloud-provider that publishes PUE for their data-centers regularly. Microsoft Azure published some data a few years ago that differentiates by continent. AWS does not release any public information regarding PUE - there only is the myth, that internal data shows PUE is in average at 1.2.
Accomplishements
That I finished right in time...
Learnings
Everything! PUE Data, typescript, using the IF, unit testing, ...
What`s next?
Some more experienced typescript-programmer should revise the plugin. Then I think it's ready, to add long-term value to the IF!