A server cluster is used to increase system stability. Each cluster is composed of individual nodes that are further classified as active and passive nodes. Active nodes are at the disposal of requests whereas passive nodes wait for active nodes to fail; they then take over their job. In a cluster all nodes can be active nodes. If an active node fails in such a cluster, its job will be taken over by another active node in addition to the latter node’s own job.

It needs to be understood that Windows clustering does not mirror the machine status of one server on another. Thus, if a node suddenly stops during a request (blue screen, power surge etc.) the backup node will not be able to continue processing the request. Rather, a cluster relies on the client noticing that a request has not been answered and repeating the request. That second request would the go to the backup node.