From: Boris Zuckerman (boris@nextpoint.com)
Date: 03/30/97-06:37:53 PM Z
Message-ID: <01BC3D41.E5F1D110@BHOME> From: Boris Zuckerman <boris@nextpoint.com> Subject: RE: UID vs Security Descriptors Date: Sun, 30 Mar 1997 19:37:53 -0500 -----Original Message----- From: Brent Callaghan [SMTP:brent@caribe-86.Eng.Sun.COM] [] ... So, I'm inviting suggestions for fixing this. There is some simple, elegant way to solve this problem isn't there ? How do NT servers handle this now ? [] Currently our servers forced to support mapping of NT users to UID/GIDs (which by the way usually a big security problem for unix clients). [] I would like to see something more open than 32 bit UNIX UID/GID scheme - probably some kind of 'printable' identities. Also it would be useful to know what is the identity (or possible set of identities) of files that are created for the current authenticated user. Note: these identities could be different from those returned during authentication. If Clients (a.k.a. Redirectors) whish to provide additional mapping to satisfy requirements of operating systems and applications, they certainly can do this. One of the noble goals in transition into NFS-4 could be substitution of the UNIX-centered philosophy of NFS for SERVER-centered philosophy. We probably should admit that there are other file systems around and the new protocol should be convenient and non-restrictive for all homogeneous Client-Server combinations.
This archive was generated by hypermail 2.1.2 : 03/04/05-01:45:43 AM Z CST