Windows 2000 Distributed File System

If your network has been running for a while, there is no doubt that your server's hard drive will be full, usually this means it's time to clear files that users no longer use. You can move files that you rarely use to offline storage, and in Windows 2000, you have another option. When the server hard disk is full: you can build a Distributed File System (DFS) tree structure. Building a DFS tree structure not only solves the problem of hard disk filling, but in fact DFS can also enhance network security, efficiency and performance.

The DFS architecture makes files and directories scattered across multiple servers look as if they are on the same server. For example, suppose your users need to read files from shared nodes on two different servers. You can make their operations simple by creating a DFS tree structure with two different shared nodes. On top of this, the user can read the DFS tree structure of a single shared node without knowing which server or shared node the file is actually hosted on.

Implementing a DFS tree structure can make it easier for members of large organizations to work. However, in many organizations, the convenience of members' work is not pronounced. Most of the time, unless the implementation of new things is conducive to the performance and security of the network, administrators are simply not too busy to do so. If you understand: Building a DFS tree structure can enhance the security, reliability, and performance performance of your network. Then, DFS may be one of the best ways to solve these problems.

First, security enhancement

First, let's take a look at security. As I mentioned earlier, the DFS tree structure incorporates a single shared node into a tree structure, making the user seem to operate in shared resources that exist on a stand-alone server. These shared nodes are only logically combined - at the same time, each shared node is reserved as a separate shared node on its own server. DFS requires that any shared node that you use to construct the DFS tree structure part uses the NTFS security attribute. Therefore, regardless of whether the shared node is read independently or through the DFS tree structure, any security attributes of the independent shared node you specify through NTFS will work at the same time.

So, since existing security attributes still apply, how does DFS enhance security? Remember that all files and folders in the DFS tree structure appear to exist on a stand-alone server from the perspective of the client? This masquerade makes it harder for hackers to track the actual location of a file. Of course, hackers will continue to explore the network and find the folder or file they need, but the longer he completes the item, the higher the risk of being captured. Many hackers get internal information from a member of the company. Since your users won't know which server the shared node's files are actually working with, dissatisfied employees can't tell the person outside the organization where the files and folders are actually stored.

Second, reliability

Now, let's look at the issues of security and fault tolerance. Windows 2000 with DFS allows you to use replicas. In a DFS tree structure, a replica is a copy of a folder that exists on a decentralized server that is scheduled to synchronize with the source server. Therefore, it is possible to have two identifiable folders on different servers.

Once you have copied the folders, you can use them when the server is offline. Suppose there is a folder that users often read, and that folder is on a server that you want to go offline for routine maintenance. Before disconnecting the server, you can change the source folder that originally pointed to the DFS tree structure to point to the folder copy. This allows the user to get a false impression that the server is still online, because the files in the folder can still be obtained in the same location as before. When the offline server is restored to the available state, the copy remains synchronized, and then you can point the DFS tree structure back to the source folder.

Third, performance

Copy can also be used to enhance server performance. You may know that if too many people try to read a single shared node, the performance of the server will suffer. To compensate, you can apply frequently used folders and copies at the same time, so you can achieve a series of network load balancing. Users still read the folder through the DFS tree structure, but the DFS tree structure connects them to the closest copy to the folder, thus distributing the load across the distributed servers and enhancing performance. For very busy shared nodes, you can cross-distribute across up to 32 scattered servers.

If I have shown you the strengths of DFS, it is time to start configuring the DFS tree structure.

IV. Installing DFS

Before you start setting up DFS, you need to know something about DFS. First, the Windows 2000 version of DFS can be implemented in two different ways. You can install DFS on a click-through server or a server farm that belongs to an Active Directory domain. Clicking on the DFS operation has some restrictions, such as not being able to read Active Directory (obviously) so it is not possible to make a copy, although some organizations may do so to enhance storage. Therefore, I highly recommend constructing a DFS tree structure based on Active Directory. The rest of this article assumes that you are installing the DFS tree structure of the Active Directory base.
The DFS client software must be running before the client reads DFS. If your client is a Windows 2000 Professional or Windows NT 4.0 workstation (Service Pack 3 or higher), then you don't need any special software. It is said that Windows Me also includes this feature, but I am not sure. Windows 98SE (Second Edition) includes a DFS client program that can be used with the Windows NT 4 version of DFS or the standalone DFS tree structure. You can also download Active Directory DFS (also known as a domain client) for Windows 95 and Windows 98 users. No other operating system supports the client's DFS structure.

V. Configuring DFS

After describing the above, now we are in the configuration process. From the Start menu, click: Programs > Administrative Tools > Distributed File System to load Microsoft's Management Console and Distributed File System Console. Next, select the New DFS Root Node command from the Actions menu. If you plan to use a pre-existing DFS root node, you can use the Show existing DFS root node option.

Windows enters the New DFS Root Node Wizard. The first screen after the introduction interface is for you to choose whether to create a new domain DFS root node or a stand-alone DFS root node. Since we are working in an Active Directory environment, select the New Domain DFS Root Node option.

The next screen asks for the domain name used by the DFS root node, the DNS name of the server hosting the DFS root node, and whether you want to create a new shared node for the DFS root node or an existing shared node. Then, you enter the name of the DFS root node and optionally add a hint that describes the DFS root node functionality. Click Next to complete the completion screen and end the wizard.

VI. Adding a Shared Root Node

Congratulations - you have built a DFS root node. You can now start adding shared nodes to DFS. Of course, unless you add an additional shared node, the DFS root node is just another shared node.

Select the DFS root node to add a shared node, and then select the New FS Link command from the Actions menu to bring up the New DFS Link dialog box. Fill in the link name, share the node, and optional comments. You can also set the client cache deferral time so that you can control when the client sees the effective time without having to see if the link is actually valid. The longer the setup time, the less likely it is to occur. However, when you may need to maintain a shutdown server you should switch the location pointed to by the link, which may cause service interruptions for a long time.

Now that you know how to configure a basic DFS tree structure, you can start to expand your knowledge. For example, you are entirely likely to implement load balancing in a DFS system.