[OpenAFS-devel] convergence of RxOSD, Extended Call Backs, Byte Range Locking, etc.

Jeffrey Altman jaltman@secure-endpoints.com
Thu, 23 Jul 2009 12:27:01 -0400


Tom Keiser wrote:
> We desperately need
> a protocol draft so that we can take this discussion over to
> afs3-standardization, and hopefully make forward progress.

I cannot echo Tom's sentiment enough.  The RXOSD implementation
as it is today is an implementation.  Even assuming that it
implemented a protocol that required no additional changes, we
would still need a protocol document that can be used to permit
kAFS, the OpenAFS Windows cache manager, Arla, etc. to implement
the functionality.

More importantly, it is impossible for anyone to rationally
evaluate the impact of accepting the RXOSD implementation into
the OpenAFS client and server without being able to visualize
the protocol that is implemented and under the semantics.  I
suspect that are good portion of the time that Tom has spent on
his white paper has been to extract from some version of the
source code the protocol.  It should not be necessary for an
implementation reviewer to perform this work.

Please spend the time to write a protocol specification for
OSD that can be analyzed.

Jeffrey Altman