Administering an ISP Installation

Fault Tolerance and Load Balancing

These features enhance the reliability of your installation by minimizing any potential interruption of service. If one server node stops working, another server node will immediately pick up the request load, with minimal disruption to users. Fault tolerance is made up of two components, failover and failback.

Failover transfers one server node to another node. Failback restores the load to the failed server node, after that node is back online.

Load balancing refers to two or more servers that support large amounts of activity by equalizing the request load among them. With Network Load Balancing, you can assign Web or FTP sites to a specific preferred server node, either manually or programmatically. In other words, a group of front-line Web servers would handle requests in the following manner: If an HTTP request comes through when one of the servers is already very busy processing requests, Network Load Balancing would direct the HTTP request to the idle server. In this way, you make sure users can access information quickly, even though a site is receiving a large number of simultaneous requests.

Note Typically, you should configure clustering to enable both fault tolerance and load balancing.