I am a Windows administrator (among other roles) and I agree with what Chris 
states here.

The way AD works (now anyways, since Windows 2000 Server) is that you establish 
a domain rooted on a DNS namespace (such as the oft-used example in MSFT 
documentation, "contoso.com".) The first such domain created also serves as the 
root of a domain tree (in the DNS sense, like contoso.com -- east.contoso.com 
-- hq.east.contoso.com, etc.), and also of a "forest" which is a collection of 
multiple domain trees. (Of course, you can have a forest with only one "tree" 
consisting of only one domain.) 

So the "forest" is actually the main security boundary; all domains within the 
forest have an automatic two-way transitive trust relationship. The individual 
domains within the forest define a replication boundary; all domain controllers 
in a given domain replicate their objects between themselves so that each is a 
full "master" in a DNS sense (excepting for 5 roles named "FSMO" [Flexible 
Single Master of Operations] for cases where you need one and only one DC to be 
a role holder.) Within a new domain, there is a default "folder" structure 
(actually, "Containers" and "OU"s - Containers can't have Group Policy applied 
to them, whereas OUs can) that have certain default objects contained within 
them. You can then go on to create your own OUs depending on your need for 
policy application and structure, and create/modify/delete objects within them.

OU's are the main way of enforcing security policy *within* a domain. So you 
can create an OU (or nested OUs) for students, for example, and have policy 
applied to the topmost student OU that will be inherited by all of the child 
OUs (if any.) Policy can be per-computer (applied to Computer objects within 
the OU) or per-user (applied to User objects within the OU.)

As you can imagine, it's a pretty deep subject, and Microsoft keeps adding to 
it as each version of Windows Server comes out. A good book on the subject 
which I find very helpful is:

Title: Active Directory, Fourth Edition
By: Brian Desmond, Joe Richards, Robbie Allen, Alistair G. Lowe-Norris
Publisher: O'Reilly Media
Formats: Print, Ebook, Safari Books Online
Print: November 2008 
Ebook: November 2008 
Pages: 864 
Print ISBN: 978-0-596-52059-5 | ISBN 10: 0-596-52059-X

HTH,
Will

-----Original Message-----
From: [email protected] [mailto:[email protected]] On 
Behalf Of Chris Francy
Sent: Tuesday, July 19, 2011 3:56 PM
To: Ski Kacoroski
Cc: [email protected]
Subject: Re: [lopsa-tech] Questions on AD domains and Bind DNS

On Mon, Jul 18, 2011 at 3:09 PM, Ski Kacoroski <[email protected]> wrote:

> 1. ad.nsd.org
> Single top level domain for staff and students.  We are concerned 
> because this will allow students to log into staff computers and see 
> resources in the entire domain.  If you have a setup like this, have 
> you seen problems with students getting into machines/resources they should 
> not?

My $0.02 is that you should avoid creating additional domains as much
as possible.   It sounds like you want a multi-domain setup mostly for
security, and I personally don't think that you get a big enough security gain 
in a multi-domain setup to offset the additional administrative costs in a K-12 
environment.

Controlling who has access to what in a single domain is relatively easy with 
policies and setting appropriate permissions.  Yes, if a tech is careless, they 
can leave something more open, and a student could access it, but that can 
happen even in a multiple domain setup.

Also consider, that while it may break policy, it is likely students have 
access to the teachers station anyway, no matter how you setup your domains.  
In my experience it is very difficult to get teachers to be really concerned 
about security.  Teachers often share their credentials with their student TA, 
and the technically inclined students who help them fix problems that cannot be 
handled by the typically understaffed K12 IT departments.

I am going to guess that you will have a maximum of about 22,000 accounts and 
somewhere between 2000-2500 computers since Northshore has ~20,000 students.  
This should be able to be handled easily by a single domain.

With a single domain you have less work to duplicate by creating policies in 
multiple locations.  You have less domain controllers to maintain.  You will 
have to put it some more effort into making sure you setup the policies and 
permissions correctly.

Even Microsoft doesn't seem to say that a separate domain creates a security 
boundary.

] http://technet.microsoft.com/en-us/library/cc756901(WS.10).aspx
] By contrast a domain is not a security boundary because within a forest ] it 
is not possible for administrators from one domain to prevent a malicious ] 
administrator from another domain from accessing data in their domain.


On Tue, Jul 19, 2011 at 9:29 AM, Dennis <[email protected]> wrote:
> I'm not super familiar with the school district, but looking at their 
> site they would seem to have tens of thousands of ldap objects so it 
> may be beneficial to segregate these out into separate domains.

For K-12 schools, I don't think this really does much.  Since 90-95% of the 
accounts will usually be student accounts, and the rest will be staff.  Even If 
there is slow browsing because of the large number of objects it isn't likely 
to be improved on the student side by breaking out the small percentage of 
staff accounts into a separate domain.

Chris Francy
_______________________________________________
Tech mailing list
[email protected]
https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators  
http://lopsa.org/

_______________________________________________
Tech mailing list
[email protected]
https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to