-----Original Message-----
Hi Carey,
I'm not sure what I need.
Here's what we want to do.
We have a domain server and a SQL server set up at our first site. On the
domain server we keep our roaming profiles including Outlook.pst files.
We're going to open new sites in about 8 months. We want to control all
sites from a central point. We want all sites to access our SQL server for
data input. We also need our managers to log in at any site and have the
same profiles that they have at their home office.
We're planning on using router-to-router VPN's between all sites.
Do I need domains at every site to have users log into? I've tried logging
into a domain through a vpn and it takes a long time (about 30 min), so I
thought that having domains that talk to each other will oliviate this
problem. I did log in with a roaming profile even though the user was a new
user and had a small profile(<1 Meg). Do I need a domain server to
establish trusts between the subnets(or just dns or wins servers)?
Right now I'm trying to find out what the best way to set this up will be.
Thanks!!!
.
Dave,
I will just comment about the need / non-need for Child
Domains. I do not necessarily like jumping in on other
people's posts but do sometimes to either add a
clarification or to point out something that I see that
has not yet been mentioned. Brian is very well-versed in
these things so I will restrict our conversation to the
Child Domain need / non-need....Brian can handle the rest
( I am quite sure! ).
To answer your question / concern about long logons over
a VPN: if Sites are set up properly your users would
authenticate against a local DC ( meaning, if a user is
located in the Miami Site, he/she would authenticate
against a DC in Miami; if a user is located in the Boston
Site, he/she would authenticate agains a DC in
Boston.....this should really cut down on that problem! ).
You see, you would *probably* have a DC ( hopefully
two! ) in each Site. Remember, they are all part of the
same domain ( in this example, yourdomain.com ). Making
use of Sites is simply a way to control Active Directory
Replication and User Logons.
Please remember that there are two types of AD
Replication: Intrasite Replication and Intersite
Replication.
Intrasite Replication is the replication that takes place
between all DCs in the same Site. So, as per my example,
all of the DCs in the New Your Site will replicate their
AD ( NTDS.dit ) with each other; all the DCs in the
Boston Site will replicate their AD with each other; etc.
etc. etc. Furthermore, please keep in mind that
replication takes place in both directions as a one-way
connection object.
Let's use New York as an example. Say that you have two
DCs in the New York Site: DCNYC01 and DCNYC02. There
would be an incoming connection object on DCNYC01
representing DCNYC02 ( that would represent the Intrasite
AD Replication from DCNYC02 to DCNYC01 ) and there would
be an incoming connection object on DCNYC02 representing
DCNYC01 ( that would represent the Intrasite AD
Replication from DCNYC01 to DCNYC02 ). Remember, this is
a one-way connection.
Furthermore, let's remember that there are three
partitions, or Naming Contexts, that are replicated
within AD: the Schema NC, the Configuration NC and the
Domain NC. The first two are replicated to all DCs in
the Forest while the Domain NC is replicated to all DCs
in a particular Domain.
How do you set this up? Well, you do not have to really
do anything *necessarily* for Intrasite Replication.
Something called the KCC ( or Knowledge Consistency
Checker ) does this all on it's own just magically. You
really do not need to do anything. Granted, you can if
you so choose. In order to do what it does the KCC makes
use of the Intrasite Topology Generator...
Intersite Replication is the replication that takes place
between Sites. So, how does this work? In each Site
there is something called a Bridgehead Server. The AD
Replication between, say, the New York Site and the
Boston Site would take place ONLY between the BHS in the
New York Site and the BHS in the Boston Site. Again,
this replication takes place in both directions as a one-
way connection object. So, you would have the incoming
connection object on, say, DCNYC02 representing DCBOS01 (
representing the AD Intersite Replication from Boston to
New York ) and you would have the incoming conenction
object on, say, DCBOS01 representing DCNYC02 (
representing the AD Intersite Replication from New York
to Boston ). Again, our friend the KCC does most of this
magically. It would use the Intersite Topology Generator
to do this. However, there is one thing that you, the
Admin, would need to do: create the Site Links and
possibly the Site Link Bridges. You simply need to
provide four pieces of information: name of Site Link and
Sites involved, Transport Protocol, Cost and Schedule.
Once you have done this the KCC has all the information
that it needs and does it's thing. Again, you can do
this manually if you choose or you could partner up with
the KCC or you can simply let the KCC does it on its
own. The choice is yours.
I could provide you a few links for this but I think that
I have covered it fairly well. Take a look at the
following two MSKB Article that describe how both WIN2000
and WINXP Clients locate a DC:
http://support.microsoft.com/default.aspx?scid=KB;en-
us;247811
http://support.microsoft.com/default.aspx?scid=kb;
[LN];314861
Also, take a look at this Article that helps to control
the Clients locating the proper DC...
http://support.microsoft.com/default.aspx?scid=kb;en-
us;306602
HTH,
Cary