|
On 8/15/2013 7:12 PM, CRPence wrote:
On 02 Aug 2013 07:20, Buck Calabro wrote:
<<SNIP>> Data access security is hard coded in the programs
(users X, Y and Z can access region 3 but not region 1; all data
in the same tables. OS security doesn't help here. <<SNIP>>
I understand that the statements are based upon the implication that
an application already had been designed as the sole means to implement
access [possibly directly to the physical file], but...
I just want to clarify... that the OS security _can_ be used, with the
database via logical views of the data and authorities to both the
physical data file and the logical file, to accomplish exactly that
[certain users access to specific data], without having to instead use
an application to control access; i.e. all accesses, irrespective of the
application that is being used [while not adopting excessive
authorities], can be prevented from undesirable access to the data in
the files.
Given this one file:
Region Customer CreditScore
1 1 100
1 2 83
2 3 90
The application has internal restrictions such that Fred can read region
1 and Ginger can read region 2. For 'application', let's simplify to a
single inquiry program written in RPG. That RPG program looks vaguely
like this:
FCREDITRATE IF E K DISK
C USER LOOKUP TABUSR TABRGN 20
...
C CUST CHAIN CREDITRATER 80
C IF REGION <> TABRGN
C* issue error
C GOTO GETCUST
C ENDIF
The database is indeed capable of supporting many logical views with
criteria that limit the visible records to specified selections. So we
could have:
CREDITFILE1 Select region 1
CREDITFILE2 select region 2
...
Is this what you were thinking of? If so, is there a way to bring all
these logical views into a single F-specification like the original
code? Where a CHAIN to an unauthorised row yields an error I can trap?
When I first puzzled this problem out I thought of dynamic SQL: use the
user/region table to generate the SELECT for the proper table and off we
go. This would be easier to write and maintain than several
F-specifications and a pile of CASE...WHEN statements to choose the
proper file to CHAIN to, but it seems to me that it will still require
significant application changes in order to implement.
When I wrote the quoted post, I didn't see a way to use the database to
restrict a subset of rows by user without application changes.
Hopefully, I'll soon see the flaw in my thinking!
--buck
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.