Load balancing is the practice of spreading work among different computers or different routers so that all computers or routers can run at optimum efficiency; by not loading down one computer with all of the work and spreading it around from computer to computer or ‘balancing the load,’ work gets done more quickly and effectively.

Firewall ISP load balancing and load balancing without BGP are all network load balancing issues that are addressed in Wikipedia, although the issue itself is not the clearest to understand. Cisco load balancing and load balance software for servers are the main issues and entries that come up in a Google search (if you decide to do one) but what does this all mean?

Load balancers are what is used to balance the load between virtual servers or computers; a load balancer is connected to a number of different servers on a ‘server farm’ and work is dispersed to the different areas of he server farm. When a client requests some information, the information passes through the load balancer which assigns the ‘work’ to that particular server; depending on the type of service or application the load balancer may send the request to any one of several virtual servers.

The great thing about some load balancers is that it can provide web site monitoring or server monitoring to some extent; there is no replacement for proper web server monitoring and good web server monitoring software but to some extent load balancers provide notifications for servers that are down, as information and requests cannot be sent to them. Unfortunately it is not as timely as regular server monitoring as server monitoring is usually on a more consistent basis and the transfer of information has to fail before the load balancer can even recognize the fact that the server is down. Server farms, caches and firewalls are all able to benefit from proper load balancing.

Load balancing can be done in a variety of ways, depending on how they are set up. Some load balancers send requests to servers with the least amount of connections already made as they are doing the least work; other load balancers send work to servers in a particular order and the first available one picks up the work. Still other load balancers process available servers as to which one picks up the request the fastest.