idra at samba.org
Thu Sep 15 15:29:51 GMT 2005
On Thu, 2005-09-15 at 08:04 -0700, Randy Turner wrote:
> Hello list,
> I had a high-level question regarding the technical goals of the
> samba4 project. In the past, my projects have included embedded
> system projects that needed CIFS/SMB client/server functionality.
> Sometimes these projects were embedded linux, and it wasn't too
> difficult to get Samba 3.0.x running on these platforms. However,
> when these embedded platforms used some other real-time operating
> system (like VxWorks), the design of Samba 3.0.x and it's OS
> assumptions made the task of porting Samba 3.0.x extremely difficult,
> and was not attempted.
> What I was hoping is that the design Samba4 was modularized and not
> as monolithic as Samba3, and that the interfaces used by Samba4 were
> not "hard-wired" to a Unix system API. I just pulled down the current
> trunk of the samba4 tree and I'm just starting to look at it, but I
> was hoping for a basic statement as to whether or not the list feels
> like the portability of the samba4 codebase is much better than Samba3.
Samba4 should be more portable than samba3.
One of the basic features of samba4 is that it does not assume you have
a posix filesystem (which is the default configuration anyway) but that
you can write your own vfs and fully exploit CIFS semantics.
Samba4 is much more modularized, event driven and can run with different
process models. Current supported models are standard (same as samba3),
single (a single process handles every service, from cifs to ldap) and
also a threaded model.
for more info look at tridge's home page where ou can find slides form
Simo Sorce - idra at samba.org
Samba Team - http://www.samba.org
Italian Site - http://samba.xsec.it
More information about the samba-technical