Mindcraft NT vs. Solaris SMB performance tests

John Lusk lusk at dg-rtp.dg.com
Fri Sep 4 13:24:36 GMT 1998


"AM" == Andrew Miller <andrew.miller at csfb.com> writes:

  AM> Serious SAMBA Admins,
  AM> Has anyone tried to reproduce the NT/Solaris SMB server comparison done
  AM> by Mindcraft:

  AM> 	http://www.mindcraft.com/whitepapers/nt4sol26exec.html

I've been doing some playing around w/Samba 1.9.18p10 here, and I can
say that Mindcraft benchmark looks pretty bogus to me.

(I was going to write a bit more and present some of my data, but I'm
not quite sure of it yet, so I'll just sit on it a bit longer.
Suffice it to say that my results are different from Mindcraft's.)

The Mindcraft writeup doesn't mention what attempts at tuning were
made on the TotalNET software, nor what kind of disks were used (a
RAID-0 stripe can make a big difference), nor what the network
configuration was.  The authors of this article may have intended it
to be a Unix-vs-NT comparison, but, in reality, it's a TotalNET-vs-NT
comparison.

In my opinion.

Which does not necessarily sync w/my employer's opinion.

John.



More information about the samba mailing list