Pending and desired OpenAFS projects

This page describes some of the projects that are currently known to be in progress or discussed within the OpenAFS development community.  This list is by no means complete.  Additional sources of information regarding on-going and future plans can be found on the road map and the Google Summer of Code pages.

Pending projects

The following are items known to be in progress, and, where available, a point of contact.
Hierarchical Storage Management Integration

Provide support for storing AFS files into a HSM system (DMF is the first targeted HSM system). Contact Ken Hornstein ( for more information. This project has been made possible in part by support from SGI.

Disconnected AFS Operations

Provides the ability to use AFS while not connected to a network. Significant work was completed during the 2008 Google Summer of Code and currently resides in the OpenAFS Git repository as part of the openafs-devel-1_5_x branch.  The current implementation is functional for read/write volume operations but cannot support hard linked objects.  There is also no method of pinning objects to the cache or applying conflict resolution policies if the content of the volume on the file changes while the disconnected client modified the same objects.  Contact Simon Wilkinson ( for more information.

FreeBSD port

Server ported. Cache manager support is not yet complete.

NetBSD port

Server ported. No timeline available.

OpenBSD port

In progress. Cache manager works but not recommended for production use.

Better Documentation

In progress. On-going projects have produced a complete set of new man pages, a replacement for the Unix Quick Start Guide, and progress on a new Administrator's Guide.  Contact Russ Allbery if you'd like to contribute

Kerberos v5 and multiple encryption types

In progress. Add Kerberos v5 and additional encryption type support. Contact Marcus Watts or Matt Benjamin

RxUDP Improvements

In progress. Your File System Inc. was awarded a SBIR grant to improve the performance characteristics of the Rx RPC mechanism.  During Phase I many implementation deficiencies were corrected and maximum throughput was raised from 60MB/second to 240MB/second. 


In progress. Implement Rx API on top of TCP/IP in order to gain performance on high speed networks and improve the behavior of AFS across firewalls and network address translators.  Ken Hornstein began the effort.  Contact Jeffrey Altman for additional information.


In progress. Implement an AFS server that exports portions of the local file system to AFS clients. Contact Derrick Brashear.

Native Windows file system support

In Progress.  Implement AFS as a Windows Network Redirector and File System Filter instead of as a virtual SMB server.  Contact Jeffrey Altman and Peter Scott for additional information.

Better server preference support

In progress.  Allow selectable continuous or on-demand server preferences in clients based on available information about Rx connection round trip times instead of classful networking approach currently used.  This project is being worked on by Jacob Thebault-Spieker as a 2009 Google Summer of Code project.  Contact Derrick Brashear (

Proposed projects

The following are items which have been proposed, but which are not yet in progress. A point of contact is included where appropriate.
IPv6 support

Make AFS support IPv6.

Multiple volume versions

Support for multiple "clones" of a single volume.

Selectable UBIK best host algorithm

Change lowest IP metric to something administrator-selectable.

Volume dump editor

Tool to allow editing volume dumps.

Partition UUIDs

Create partition UUIDs such that the vldb can be rapidly updated if a disk is moved from one machine to another. Contact Derrick Brashear (

Directory format extensions to support Unicode filenames

Protection Server database extensions and RPCs

To support multiple security class authentication name to AFS ID associations