Composite Container-based Application Specification

Problem Statement

Currently there is no standard mechanism to define a composite multi-container application or composite service composed of aggregate pre-defined building blocks spanning multiple hosts and clustered deployments. In addition, the associated metadata and artifact management requires separate processes outside the context of the application itself.

What is Nulecule?

It's a made-up word meaning "the mother of all atomic particles" pronounced NOO-le-kyul, like "molecule".

Nulecule defines a pattern and model for packaging complex multi-container applications, referencing all their dependencies, including orchestration metadata in a container image for building, deploying, monitoring, and active management.

Nulecule specification enables complex applications to be defined, packaged and distributed using standard container technologies. The resulting container includes dependencies while supporting multiple orchestration providers and ability to specify resource requirements. The Nulecule specification also supports aggregation of multiple composite applications. The Nulecule specification is container and orchestration agnostic, enabling the use of any container and orchestration engine.

Glossary of terms

Nulecule Specification Highlights

• Application description and context maintained within a single container through extensible metadata
• Composable definition of complex applications through inheritance and composition of containers into a single, standards-based, portable description.
• Simplified dependency management for the most complex applications through a directed graph to reflect relationships.
• Container and orchestration engine agnostic, enabling the use of any container technology and/or orchestration technology

“The Big Picture”

Deployment User Experience

Here's an example using the atomicapp reference implementation with a kubernetes provider.

Option 1: interactive

Run the image. You will be prompted to override defaults

[sudo] atomic run projectatomic/helloapache

Option 2: unattended

1. Create file answers.conf with these contents:

    [general]
    provider = kubernetes
   
    [helloapache-app]
    image = centos/httpd # optional: choose a different image
    hostport = 80        # optional: choose a different port to expose
   

2. Run the application from the current working directory

    $ [sudo] atomic run projectatomic/helloapache
    ...
    helloapache
   

Option 3: install and run

You may want to download the application, review, edit the answerfile then run.

1. Download the application files using atomic install

    [sudo] atomic install projectatomic/helloapache
   

2. Rename answers.conf.sample

    mv answers.conf.sample answers.conf
   

3. Edit answers.conf, review files if desired and run

    $ [sudo] atomic run projectatomic/helloapache
    ...
    helloapache
   

Developer User Experience

1. Develop an architecture to define how the services are connected and exposed.
2. Create Dockerfiles and artifacts to package services as container images.
3. Create provider files, e.g. for kubernetes pod, service, replication controller files
4. Reviews example templates of the container application specification
5. Builds the deployment container describing the application.
6. Pushes all container images to a registy.

Implementations

This is only a specification. Implementations may be written in any language. See implementation guide for more details.

Reference implementation https://github.com/projectatomic/atomicapp

Developer tooling

Developer implementation provides tooling to help developers quickly package several containers as a unit. It may be as simple as generating a template to start from or as complex as a GUI to develop and provide graphical representation of the target deployment.

Contributing

Please review the contributing guidelines before submitting pull requests.