Different Types of Clusters

There are two types of cluster techniques in the Microsoft server family. An administrator can use these two separately, or in a combination. One does not rule out the other, but both can instead work together to help an organization provide service for its partners and customers, or internally. Windows clustering technologies include those discussed in the following sections.

Network Load Balancing (NLB)

NLB service balances the incoming Internet Protocol (IP) requests among a cluster of up to 32 servers. This way it helps in solving both the scalability and the availability problems. Several servers can share the workload, and if one server fails, the others handle its work without the users noticing. This technique is often used in large Web farms or terminal server farms, for instance. You will learn more about this technique later in this chapter in the "Network Load Balancing Overview" section. NLB creates a virtual server so that clients only have to know the name or IP address of this server, no matter which server actually handles their request.

Microsoft Cluster Service (MSCS)

This service helps primarily in solving the availability problem. A cluster service works by having two or more nodes in its cluster. If one node fails, the other (or others) takes over. This is known as failover technique. So when do you use this? If you have a database like Microsoft SQL Server, you can use this technique to provide availability for the users and make sure that data has not been compromised during the failover. You can also use this with Microsoft Exchange Server. Again, the clients only have to know the IP address or the name of the cluster and not bother with the individual servers in the cluster.

Combining the Two

As mentioned earlier, these two technologies can work together. You can deploy NLB across the front-end Web servers and use MSCS at the back end to cluster the SQL Server providing the Web pages with data (see Figure 3-1). This way you make sure that both scalability and availability are as high as possible.

Figure 3-1: Combining MSCS and NLB

When to Use What

Now that you have learned the basics about clustering, you might be wondering when to use what technique. Table 3-1 helps you with your decision.

Table 3-1: Overview of When Each Clustering Technique Is Suitable for Use
Application	NLB	MSCS	Benefits
Web server farm	X		Enhances scalability and availability
Terminal server farm	X		Enhances scalability and availability
File/Print server		X	Enhances availability and data consistency after failover
Database server		X	Enhances availability and data consistency after failover
Messaging server		X	Enhances availability and data consistency after failover
E-commerce	X	X	Enhances scalability and availability

Note

But there is still one problem. With a cluster of 32 servers or perhaps several clusters, how on earth do you as an administrator handle the maintenance of such a number of computers without going completely mad? What we are talking about here is manageability. Well, it might sound difficult, but luckily tools are available to help you out. These tools present the cluster as one single point of control. And yes, here you definitely want a single point so as not to go crazy. Some tools are integrated in the operating system, but Microsoft has also developed a server, called Application Center, to help the administrator with these tasks. Application Center will be presented, along with some of these tools, later in this chapter.