Roadmap

The short term goal is to stabilize the Linux kernel client and userspace daemons to facilitate testing and debugging the file system.  We hope to have the system usable (for non-critical applications) by the end of 2009 in a single-mds configuration.  Even then, however, we would not recommend going without backups for any important data, or deploying in situations where availability is critical.

Notable upcoming architectural extensions and changes include:

• Strong scalable security. There is a proof of concept implementation based on this paper, but it needs to be brought up to date with the current code base. Node and user authentication, encryption, etc.
• Content-addressable storage. Name objects based on a hash of their content, allowing files to be transparent ‘deduplicated’ within the storage system. A digester will process new files (or follow some other policy).