[jcifs] SMB URL @ CIFS Conference
Christopher R. Hertel
crh at ubiqx.mn.org
Wed Aug 22 13:10:35 EST 2001
Yes/no.
It's not just CIFS vs. SMB but native TCP vs NBT *and* all the supporting
protocols that go along with that. Browsing, name resolution, and
authentication are all different with port 445 semantics (the new stuff).
So, for example, what does smb://<w2kdomain>/ mean? By logical extension,
it would mean "give me the browse list for the Windows 2K domain".
Getting that list, however, is now an Active Directory call (I think)
rather than a call to a local browser. So, if we overload smb:// to mean
*both* NBT and non-NBT transports we need to demonstrate that mechanisms
exist to resolve a name (eg. smb://service/) to any of the below:
+ It's a NetBIOS name
- it's a workgroup name
- or it's a server name
+ or it's a partially qualified DNS name
- so it's a server running NBT services
- or it's a server running non-NBT services
- or it's the name of a W2K domain and we want the browse list
See the problem? We've added more permutations which means more testing
for the underlying protocol implementation to handle.
Chris -)-----
On Tue, Aug 21, 2001 at 07:02:42PM -0400, Allen, Michael B (RSCH) wrote:
>
> > -----Original Message-----
> > From: James Nord [SMTP:teilo at teilo.net]
> >
> > What can a CIFS client do that a SMB client can not (and vice versa)
> >
> Nothing. This is just a terminology issue. I'm of the opinion that SMB
> is the technical name for the type of transport used with the CIFS
> protocol -- Transactions, RAP functions, Named Pipes, etc are all
> implemented over SMB. However, CIFS is the newer more all encompassing
> name for the protocol as a whole and people are starting to use the cifs://
> prefix. The whole thing is a little unfortunate because supporting both
> interchangably creates confusion and programmatic problems (e.g. create
> an object with smb:// and then compare it with an object created with
> cifs:// ?). Also, implementations will no doubt choose to attach meaning to
> one or the other (e.g. Chris has advocated cifs:// be used to indicate
> NetBIOSless transport).
>
> Mike
>
>
--
Samba Team -- http://www.samba.org/ -)----- Christopher R. Hertel
jCIFS Team -- http://jcifs.samba.org/ -)----- ubiqx development, uninq.
ubiqx Team -- http://www.ubiqx.org/ -)----- crh at ubiqx.mn.org
More information about the jcifs
mailing list