[Discussion]: "The problem with the carbon-aware-sdk"

[PhilippMarla]

My issue about the carbon-aware-sdk is, that it won‘t save carbon emissions!

The SDK will shift power consumption in time or in location and both without involving the transmission system operators (TSO), which are the operators of the electricity grids. The TSO plans the electricity supply at least one day in advance for each quarter of an hour in a very exact manner. Based on the previous consumption of each datacenter, the TSO will do a forecast of the energy demand. This is the base of the power dispatch plan, where each power plant will get its delivery request.
A significant amount of power consumption, which will be shifted using the SDK, will disrupt this detailed plan. The TSO is now forced to act rapidly by utilizing quickly controllable energy source. As the main renewable energy source solar, wind and water rely on the weather, we can not control them and above all we can not increase them. This means a quick increase in energy demand usually leads to an increase of carbon-based power, as they are better controllable.
As above explained the SDK has no positive influence on the carbon emission when it comes to significant power shifts. And what is about small changes?
It is just the same, there won’t be a positive effect for carbon emission. The only difference is, that you can not really measure it, because the amounts a neglectable.
As the carbon efficiency benefits presented in the different projects of the Carbon Hack 22 are based only on the numbers produced by the sdk, it is only a “self fulfilling prophecy” which lacks a real physical measurement, especially the short term controlled energy shifts are not included.
In the attached slides I tried to explain the mechanism with some pictures. I hope it helps.

I'm looking forward to some constructive discussions about that topic.
Please feel free to contact me, ask questions or what ever.

As a solution for the mentioned problem my suggestion is, that we should think about involving the TSOs. Because they are the only parts in the energy delivering system that can reliably optimize the dispatch of renewable energy.
I thought about a different approach of an SDK, that reports flexibilities of calculation loads. These will be adapted and accumulated to power demand flexibilities within each datacenter. Then the datacenters can report their flexibilities to the TSO. When it comes to the dispatch planning these information can be used to optimize the utilization of renewable energies.
And there is e very nice side effect. The reported flexibilities can also be used for redispatch purpose, which safes some money.

Here is the Link to the Slack-Thread about that topic
(There are also some slides attached)

Code of Conduct

• I agree to follow this project's Code of Conduct

Feature Commitment

• I commit to contributing this feature as a PR and working with the GSF to merge this feature into the Carbon Aware SDK.

[github-actions]

This issue has not had any activity in 120 days. Please review this issue and ensure it is still relevant. If no more activity is detected on this issue for the next 20 days, it will be closed automatically.

[github-actions]

This issue has not had any activity for too long. If you believe this issue has been closed in error, please contact an administrator to re-open, or if absolutly relevant and necessary, create a new issue referencing this one.