Summary of Exam Objectives

The two high-availability technologies included in Windows Server 2003 are server clusters and NLB clusters. Server clusters provide failover support for applications, primarily databases. NLB provides parallel servicing for TCP/IP-based network applications, such as Web servers, firewalls, and DNS servers.

Server clusters use specialized high-end, high-capacity hardware, and shared storage components to increase availability. An application running on a server cluster runs as a single instance of the application, although multiple separate and unrelated instances may be running on a server cluster. NLB can be used on generalized or low-end hardware and requires nothing extra to implement, except for an optional second network interface adapter. Each host in an NLB cluster has separate copies of the applications, which are unaware of each other’s existence.

A server cluster can be configured in one of three ways depending on its intended purpose and geographic operating environment. It can also be configured to follow a certain pattern of behavior when a failure is encountered. NLB clusters are fundamentally configured identically each time an NLB cluster is created, except for the way the cluster handles client traffic and internal communications. NLB cluster error-control behavior is limited to excluding a failed or improperly configured host from the cluster.

Server clusters and NLB cannot coexist on the same hardware. However, the two technologies can be used together to form an extremely reliable foundation for fast, reliable service. NLB is often used to create a front-end structure to handle incoming client requests. The hosts in an NLB cluster then make requests of applications residing on a back-end server cluster. This structure allows efficient processing of large numbers of client requests and reliable availability of data.