Modern software development must become climate-friendly and resource-saving. However, we don't have to reinvent the wheel, as Green Software Development design patterns already exist. After the “usual suspects” such as maximizing server utilization, minimizing memory and processor utilization and dynamic scaling, a paradigm shift is needed.

Similar to the energy transition, the use of resources must be aligned with the supply of resources. This is the principle of carbon aware computing.

One option is to shift plannable loads (tasks, jobs, batch processes) to times with a high proportion of renewable energy in the power grid (time-shifting).

A different strategy is required for on-demand loads and scaling. Here, the application must adapt the supply and its functionality to the available resources. This process is called demand shaping.

This talk will show how these patterns can be implemented in the cloud with Azure and in Kubernetes and what options an application has to “shed” the load. All aspects will be addressed: How does scaling work, how can you limit it dynamically based on the CO2 intensity of the power grid and where do you get the prediction data from. And the whole thing with lots of live demos.