At a surface level, this comparison is valid because the Yoke core CLI offers functionality very similar to Helm. The key difference, however, lies in the type of packages it manages. Helm uses charts (collections of templated YAML files that, given some values, output resources), while Yoke uses flights (programs compiled to WebAssembly that read input from stdin and write resources to stdout).

However, as a project, Yoke believes that client-side package management is only a stepping stone toward server-side package management.

Client-side package management is not fully aligned with the ethos of Kubernetes. Kubernetes is designed to be extended with APIs that are created, validated, and authorized by the control plane. By deploying on the client side, we forgo many of the capabilities Kubernetes offers, often to our detriment.

In the past year, we have seen a shift toward server-side solutions, with new projects emerging to enable resource and package abstractions built directly on Kubernetes. Examples include KRO, Crossplane Functions, Koreo, and others.

It should come as no surprise, then, that the Yoke project has its own server-side solution for this purpose: the Air Traffic Controller (ATC).

Similar to KRO, the ATC enables server-side package management, but with the same key difference that distinguishes the Yoke CLI from Helm: there's no YAML—just code.

How Does It Work?

1. Define a Custom Resource Definition (CRD): Write a CRD type in your code.

2. Write a Program (Yoke Flight): Create a program that reads an instance of the custom resource from stdin and outputs the desired resources to stdout.

3. Create an Airway: Use an Airway (a custom resource included with the ATC) to define your new CRD and associate it with the program you wrote.

4. Deploy Packages: Use your newly created custom resource to deploy packages via the Kubernetes API.

With this approach, we encapsulate all of our Kubernetes application logic into a single program without the need to build a custom operator. The only logic required is the transformation of our new custom API into a set of Kubernetes resources. This method retains all the advantages of a comprehensive development environment, including type safety, ease of testing, IntelliSense, and the full range of features you would expect from a modern coding environment.

For more information, visit the docs: https://yokecd.github.io/docs/airtrafficcontroller/atc/ or follow along with the examples https://github.com/yokecd/examples written in Go.

We’d love to hear your thoughts and feedback on Yoke’s Air Traffic Controller! Feel free to share your ideas, use cases, or any challenges you encounter. Let us know what you think!