Note: Many topics at this site are reduced versions of the text in "The Encyclopedia of Networking and Telecommunications." Search results will not be as extensive as a search of the book's CD-ROM.

Load balancing is the process of distributing some load across multiple links, servers, processors, or other devices in order to improve performance and overcome deficiencies in existing equipment. This topic is primarily about server load balancing, which is the process of distributing requests from clients across a group of servers.

Another type of load balancing takes place in aggregated communication links, where network traffic is distributed across two or more links that appear as a single link with high bandwidth (and redundancy). See "Link Aggregation."

Still another type of load balancing occurs in multiprocessor systems, where processing loads are distributed across multiple processors in the same computer, or across an array of computers that are configured into a multiprocessing cluster. For example, a single complex task may be broken into pieces that can be simultaneously processed on different processors. See "Multiprocessing" for more information.

Server load balancing is a hot topic when you consider that most Web sites are overworked and old server configurations are unlikely to handle the load. Multiple servers are required, some that perform specialized tasks, so the job of a server load balancer is to receive incoming traffic and send that traffic to available servers or servers that are most appropriate for a particular task. These servers are represented by a virtual IP address. Traffic destined for the IP address is intercepted by load balancers and distributed to servers in a server farm or cluster. Also see "InfiniBand" for information about switching technology for servers.

Figure L-6 (see book) illustrates a typical load-balancing solution for a Web site.