Server clusters provide high availability and fail-over functions for enterprise applications and servers. For enterprises that have not already begun creating high availability environments, server clustering, using the software provided with the operating systems, can offer the first step towards continuous availability.

Server Clustering

When an application or server fails, clustering software will restart the failed applications on an available fail-over server. By using application aware agents, an internal application error can be detected even if the application continues to appear to operate. For example, an agent will catch a Web server that continues responding to requests, but only provides "404" error messages.

Depending on the software and the configuration used, clustering can be equipment intensive, requiring a one-to-one fail-over to primary server in the worst case. Larger environments can utilize distributed clusters, where a single server may act as a backup unit for many different applications or multiple servers may provide fail-over resources to all servers in a cluster.

Windows Server 2003 (and dating back to Windows NT), Linux, and most Unix variants include clustering capabilities. Optional or third party packages can provide advanced...