In Windows NT 4.0, Microsoft provided an add-on product called Distributed File System (DFS) that allowed physically separate network file resources to be grouped together and accessed as if they were a single logical structure. The product, which was a free download, failed to make a great impact with network administrators and went largely unnoticed. With Windows 2000, DFS is included with the OS and provides a number of new functions. The tool for managing the DFS structure has been improved, and wizards serve to make setup an easy task.
DFS is a service that gives administrators a way to provide users with simple access to increasingly distributed amounts of data. In this article, I will look at some of the features of DFS and how to create a DFS tree in Windows 2000.
|DFS in a Heterogeneous Environment|
The functionality of DFS is not just limited to Microsoft Operating systems. For instance, if the server hosting the DFS root has access to a NetWare server through client or gateway software, directories on the NetWare server can be added to the DFS tree. This is a major advantage to administrators managing data in a heterogeneous environment.[end]
DFS file structures can be accessed from any workstation that is running the DFS client software. This software is included with Windows 98, Windows NT 4.0, and Windows 2000. A downloadable client is available for systems running Windows 95. To take full advantage of the fault tolerance capabilities of DFS, the updated Active Directory Client Extensions must be installed for the respective client platforms.
What Is DFS?
DFS provides the ability to create a single logical directory tree from different areas of data. The data included in a DFS tree can be in any location accessible from the computer acting as the DFS root. In other words, the data can be on the same partition, disk, or server, or on a completely different server. As far as DFS is concerned, it makes no difference. A DFS tree appears as one contiguous directory structure, regardless of the logical or physical location of the data.
After the DFS root is created, links to directories can be added or removed to construct the single logical directory structure. The DFS tree can be navigated using standard file utilities such as Windows Explorer. Unless users are made aware of the fact that the data is being accessed from different locations, they will not realize that they are using a DFS system at all.
DFS trees can be used with both FAT and NTFS partitions. If you do use NTFS, the inclusion of a file or directory in a DFS structure has no effect on security permissions.
There are two types of DFS:
- Stand-alone DFS--Refers to a DFS tree that is hosted on a single physical server, and is accessed by connecting to a DFS share point on that server. DFS configuration information is stored in the server's Registry. Stand-alone DFS provides no fault tolerance. If the server hosting the DFS root should go down, users will no longer be able to access their data unless they explicitly know where the data is stored.
- Domain DFS--Provides more functionality, including features such as replication and load-balancing capabilities. Domain DFS information is stored in Active Directory. A domain member server must act as the host for the DFS tree. By storing the domain DFS configuration in Active Directory, the server-centric nature of stand-alone DFS is removed, enabling the administrator to create DFS root replicas. If a server were to go down, users would be redirected to a DFS root replica and could continue to access the DFS tree.
|DFS Disk Space Reports|
When a DFS share is accessed, the amount of free disk space on the drive is reported for the drive that hosts the DFS root. This amount will often differ from the amount of disk space available through different areas of the DFS structure. As an administrator, this change is easy to account for, but it can be confusing for users.