Release Status¶

Arvados is currently (April 2014) ready for adventurous users to try out, but not ready for day to day production use. If you are interested in helping to design and improve Arvados, you can do that now.

Some components have been implemented, and we are running them in our own test/production environments.

• REST API server
• Client-side pipeline manager
• Workbench web application

Some components currently exist only as prototypes from the Free Factories predecessor project. Porting these to Arvados is one of our current development efforts.

• Keep server
• Crunch MapReduce job manager

Some components have not been implemented yet. We think it's important to describe them to establish scope. We also hope to attract early feedback.

• Lightning
• Compact genome format

APIs are in active development.

• REST API is getting stable but isn't at 1.0 yet. A more powerful graph API is also in development.
• Keep client API is stable, but system-facing API (e.g., related to garbage collection) are still in development.
• Crunch API (job task environment) is mostly stable, but containerizing workers/tasks may motivate some changes.

SDKs are in active development.

• Rails SDK (an ActiveRecord-like interface) is not yet packaged separately from the Workbench app.
• Ruby SDK is stable but lacks a native Keep client.
• Python SDK (based on the Google API client) is stable but should gain an alternative (less verbose) calling syntax and better integration of data APIs with Collections.
• Command line tool should gain an alternative (less verbose, less quoting-intensive) calling syntax.

Documentation is in active development.

• Tutorials and examples are continuously improved as APIs improve and stabilize
• Install docs are terse but functional
• API docs need to be filled in (many models are currently documented only as source code in /services/api/)
• Admin docs are a placeholder