Al,

Version 1 of your proposal is exactly what I have done in the past, and
works quite well (within some defined boundaries).  Account managers were
delegated very minimum rights over AD (such as unlock account reset password
etc), and everything else was done via a "tool" (was client-server at the
time).  Account Managers would use the tool to create / manage a user
account, and the back-end applied rules to where things occurred.  The
advantage of this was that there was a business logic layer in the
background that forced certain things to occur (such as creating the home
drive / Exchange mailbox on the logically closest server, without the
creator needing to know what was going on).  The primary disadvantage was
the maintenance overhead of such a "tool".

Some Other advantages:
- Business Rules automatically defined which groups, resources users had
access to (based on a defined "profile", or location, or job function)
- Approval Paths (workflow) could be placed over the top to send an approval
request upwards if a certain level of network access was requested (eg
Administrator access). Link this in with outlook forms and the Exchange
Workflow engine and you have some pretty neat abilities.
- Granular security.  The tool (through security settings or business logic)
defined exactly what requestors could do, based on their relative ability
(or security clearance).  Functionality within the tool can be enhanced (or
completely removed) based on this security "profile".
- Requests can be scheduled !! This was a biggie at a former employer, since
comms links were heavily utilised during the day.  Requests for user
movement / mailbox movement / creation could be scheduled to occur after
hours.  The system simply delayed the task until a certain time, then
automatically processed the request.  When complete, either a notification
through the app was sent to the helpdesk person, or an email sent directly
to the requestor.
- Because there is business logic / rules applied, things always end up the
same way (capitalisation, phone number formats, email addresses etc)
- The business logic / processes could be changed easily without having to
go through a notification / retraining process.  The helpdesk etc just
filled out the form the same way, but the back-end process did other things.
- Logging.  Since the "tool" was client server, extensive logging can be
performed, and NT Authentication used to log exactly who made the request
etc.  Made it a lot easier looking for inappropriate changes to the network
(although the number was heavily reduced as no-one had access to do
anything).
- Multiple Interfaces.  We ended up with a web-based, standalone app,
Outlook Forms, and Email request methods.  Allows a lot of flexibility.
- Incremental Functionality.  Since there is an app (web or otherwise),
functionality can be incrementally added as you put functionality into the
system.  Start off small (like user creation), and increase the tools
functionality as you go.  Makes it easier to implement rather than biting
off a huge chunk in one go.
- Removes dependence on 2nd / 3rd level support.  Since 2nd / 3rd level
support typically are required for some of the higher level functions, you
can allow the app to do this without increasing the access for your helpdesk
staff.  For example, our user creations required the home drive creation to
be sent to 2nd level (due to NTFS permissions settings).  Once the tool was
doing the user creation, 2nd level no longer needed to be involved. It
released them for other / higher-value tasks.

Disadvantages
- Tool Availability.  If the tool isn't available (server crash / app crash
etc), people cant do things.
- Account Access.  Since the back-end process is doing everything, this
account may need very high levels of access (IT Security may have an issue
with that).

Haven't played with MIIS much, but I presume it could do a lot of this stuff
through the provisioning engine. The EA / Quest tools as you mentioned have
a lot of this sort of functionality as well.

Oops, that was a bit of a ramble, apologies everyone.

Glenn
 

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Mulnick, Al
Sent: Wednesday, 29 September 2004 3:06 AM
To: [EMAIL PROTECTED]
Subject: RE: [ActiveDir] How to take away the password never expirers chec k
box right?

Some alternatives to your thinking might be worth discussing:

One alternative would be to only grant them the bare minimum rights they
need to do their jobs.  An optimal way to do this would be to have a
centralized account lifecycle management tool that you can use to prevent
admins from going outside the processes defined.  Creation/Deletion and
possibly modifications would be handled by this central tool (think web
page).  From there, the local admins would have "helpdesk" rights via the
ADUC.  The centralized tool is still self-service for the admins to manage
their OU's, it's just that they don't have the direct rights.  I figure
they'd need to be able to add/remove computer objects, create OU's, and
reset passwords.  Everything else could be done via a central tool.  The
tool concept I'm thinking of would be a MIIS like tool permissioned for
their use.

Another alternative would be to enforce after the fact.  You said it's a
policy, so by nature if you're an admin you should be following policy.  If
you're not, then a) we need to know about (reporting) and b) you need to be
corrected according to company guidelines and policies.  This setup infers
more layer-7 influence at the mgmt level but you have to collect the
information and trigger the alerting. Many monitoring systems can monitor
event logs or you could home-brew something I'm sure.  A script that checks
all user objects would work as well, right? It's just that it's after the
fact vs. prevention.

That gives you three options to meet your overall goals.  To meet your
original goal of removing that right, you'd have to either go with the Quest
type tool else remove their rights from the useraccountcontrol attribute.
That's difficult since they can create objects and could theoretically just
add those rights back meaning you'd have to work harder to prevent that and
then monitor that process etc., right?

Al

 


List info   : http://www.activedir.org/mail_list.htm
List FAQ    : http://www.activedir.org/list_faq.htm
List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/

Reply via email to