Administering an ISP Installation

Building the Third Tier

To achieve high availability on the third tier, you need to install Cluster Service. To store all of the back-end data that makes up the third tier, most enterprise customers require a single, high-end symmetric multiprocessing (SMP) server (for example, an eight-processor SMP with 4 gigabytes of RAM) running SQL Server 7.0. Cluster Service can handle mission-critical database management, file and intranet data sharing, messaging, and general business applications.

Instead of relying on one high-end Windows 2000 Server, it’s best to add a second server for failover. Joined together, these two servers offer higher availability for incoming requests and simplify the task of managing data and applications. Cluster Service not only allows you to connect two servers into a cluster, but can also automatically detect and recover from server or application failures. In addition, it can juggle server workloads, in order to allow for planned maintenance without downtime.

Cluster Service has a built-in mechanism that can handle server failure. If Cluster Service detects that a server has failed, it automatically transfers ownership of resources (such as disk drives and IP addresses) from a failed server to a functioning one. It then restarts the failed server’s workload on the functioning server. The entire process, from detecting a failure to restarting the other server, typically takes under a minute.

If an individual application fails (but the server does not), Cluster Service will typically try to restart the application on the same server. If that fails, Cluster Service moves the application’s resources and restarts the application on the other server. For detailed information about Cluster Service, see the Windows 2000 Advanced Server online documentation.