Why a DFS Consolidation root?
DFS Consolidation roots are a way geting rid of old servers while keeping the name functionality of the old servers.
Consider the company Acme Computing. They are an old company with lots of old file server all over the place. New IT management has decided that all shared data should be available on the same file server. So this is kind of easy, lets just copy all the data to the new server, but wait.. There are old systems that have hardcoded paths to the old servername, this would break them. Some people might suggest just inheriting the share names and add the old names as alternative names of the server. This might work with the smaller companies but all you have done is complicated the fileserver even more. If we look at DFS, we can use an DFS consolidation root to trick the systems that the old paths are alive too. You cant configure a domainbased DFS namespace as a Consolidation root only standalone roots.
What happens (simplified)?
- The client application/system requests the file
\\oldfile\share\important\path\here\file.xml - The client looks up which computer to request the file from and gets the DFS consolidation root
- The client asks the DFS server for the file
\\oldfile\share\important\path\here\file.xml - The DFS server will check if there is an DFS root called #oldfile on the server and then rewrite the request.
- The DFS server rewrites the request as
\\dfsserver\#oldfile\share\important\path\here\file.xml - The DFS server returns the first dfslink in the path to the client, in this case
\\dfsserver\#oldfile\share=>\\filecluster\commonsystemX - The client will now request the file from the \fileclustercommon share instead combined with the orignial you get
\\filecluster\commonsystem\important\path\file.xml - The client application will now get the data it requested, even though the old file server is gone and the structure rebuilt.
A small blog entry on how to install a test DFS Consolidation root on a windows server.