SAMBA vs NFS V3

[David Collier-Brown]

Patrik Linder <patrik.linder at dimension.se> wrote:
> I want to have a comparsion with NFS v3 that shows some statistics with
> both attribute intensive and data intensive clients.

  That's mildly tricky: the classical PC client doesn't have a behavior
that is close to either of the two, as traditionally defined in the
nfsstone/nfhfsstone/laddis world.
  The usual SMB benchmark is netbench, which had a different operation
mix than either. 	

  HOWEVER: that doen't make it impossible by any means.
If anyone has examples of pc clients activity recorded from nfsstat,
then one can load nhfstone with that mix and measure server performance.
  It's more work to be able to compare netbench with anything else
(including,
alas, netbench run at a different site) In principle, if one had a
script
that produced the same operation mix as netbench (it's published), one 
could make a deterministic smb test and an equivalent lmbench test, then
compare them.

  I'll volunteer to help: send me mail, folks!

--dave  
-- 
David Collier-Brown,  | Always do right. This will gratify some people
185 Ellerslie Ave.,   | and astonish the rest.        -- Mark Twain
Willowdale, Ontario   | http://java.science.yorku.ca/~davecb
Work: (905) 477-0437 Home: (416) 223-8968 Email: davecb at canada.sun.com