[OpenAFS] OpenAFS speed - some benchmarks

[Russ Allbery]

Rodney M Dyer <rmdyer@uncc.edu> writes:

> Comparing AFS with NFS is just too complicated, the feature set is too
> different.  If you think you can, you're probably smoking the wrong
> stuff.  Maybe I'm wrong, but I think not one AFS user is going to throw
> away all the good stuff just to get more speed.

I doubt anyone with a large AFS infrastructure is going to throw it away,
but there are some reasons why people would be looking at NFSv4.  One of
the reasons that came up here recently was decentralized administration.
Right now, all AFS servers have to be highly trusted because they all
share a powerful key.  I believe this is not the case with NFSv4, which
would allow a department with a grant to throw up their own NFSv4 server,
have referrals inserted at the right place in a global namespace, and be
off and running without compromising the security of the whole
infrastructure if their system is broken into.

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>