This simple demo runs on your own machine and makes it easy to experiment with microscaling and see containers running locally on your machine. Full instructions for running the demo are at app.force12.io.
The easiest way to run Microscaling-in-a-box is to follow the instructions. The docker run
command
pulls the latest image of this code from Docker hub.
If you want to build and run your own version locally:
- Clone this repo
- Build the code as a linux executable (since it runs inside a linux container):
GOOS=linux go build -o force12 .
- Build your own version of the container image, and give it a tag:
docker build -t <your-tag-name> .
- Specify
-it <your-tag-name>
instead of-it force12io/force12:latest
ondocker run
so that it picks up your version of the image
We've got lots of ideas for improvements - here are a few headlines:
- We're planning to add support for a number of different schedulers. If there's one you'd particularly like to see us support, please let us know.
- We can improve performance by parallelizing requests to the Docker remote API.
- In this demo we simply randomize demand, but we'll add support for real-work demand models.
- The state API is currently hard-coded to our demo priority1 & priority2 tasks. We'll make it much more generic to support any task more complex scenarios.
We'd love to hear from you at hello@force12.io or on Twitter at @force12io. And we welcome new issues or pull requests!