Clustering

Are you preparing for IT certification? With practice questions, study notes, interactive quizzes, tips and technical articles, uCertify PrepKits ensure that you get a solid grasp of core technical concepts to ace your certification exam in first attempt.

Clustering

Rating:

A cluster is a group of two or more computers (servers) connected to provide fault tolerance and load balancing. It is dedicated to run a specific application. Each server in a cluster is known as a node. The failover and failback capabilities of a cluster bring the application downtime to zero.

Note: Server clustering is intended to provide high availability for applications and not for data.

Failover

In the cluster, each node or computer runs the same critical application. In case one computer fails, the other computers detect the failure and take charge immediately. This phenomenon is called failover.

Failback

When the failed node returns back to the network, other nodes take notice and the cluster begins to use the restored node again. This phenomenon is called failback.

Types of Clusters

Windows Server 2003 supports two types of clusters:

• Server clusters
• Network Load Balancing (NLB)

Server Clusters

In server clusters, all nodes are connected to a common data set, such as a storage area network. All nodes have access to the same application data. Any of these nodes can process a request from a client at any time. Nodes can be configured as either active or passive. Only an active node can process requests from clients. In the event of a failure of the active node, the passive node takes charge and becomes active. Otherwise, the passive node remains idle.



Server clusters are created for running applications that have frequently changing data sets and have long-running in-memory states. The applications such as database servers, e-mail and messaging servers, and file and print services can be included in server clusters.

A server cluster is treated as a single destination for a client. It has its own name and IP address. This address is different from the individual IP addresses of the servers in the cluster. Hence, when any server fails in the cluster, the passive server becomes active. Clients send their requests to the server cluster address. Therefore, this change over does not affect the functionality of the cluster.

Windows Server 2003 supports eight nodes in a cluster. However, Windows 2000 Server supports only two nodes in a cluster.

Network Load Balancing

Network Load Balancing (NLB) is a type of clustering. It is used to provide high availability and reliability of the application servers. NLB is configured for the applications that rarely change and that have very small data sets. Web servers, FTP servers, VPN servers are the areas where NLB can be used successfully.



In the NLB cluster, all nodes are active and have separate identical data sets. Multiple servers (or nodes) are used to distribute the load of processing data. Clients send the requests to the cluster, and then the clustering software distributes incoming client requests among the nodes. If a node fails, the clients' requests are served by other nodes. Network Load Balancing is highly scaleable. Both Windows 2003 and Windows 2000 operating systems support NLB clusters of up to thirty-two nodes.

Rating: