Todays network is under higher demand than ever before, each day handling ever growing and more complex business and Web applications. A single server will not be able to handle the load demand required; multiple servers are needed and must be used to meet the demand. For ease of use, a single IP address or domain name must be used; however, if several servers are required, this adds complexity. Using a load balancer can ease the complexity of the network setup.

With the continued development of web applications, virtualization, cloud computing and more, a load balancer has become an integral piece of networking equipment. With the use of a load balancer, a single service can be provided from multiple servers. A load balancer runs on Layers 4 through 7 of the OSI Model, and some of the common protocols that are usually load balanced are HTTP, IRC, FTP, NNTP and DNS. When in use, a load balancer is placed in front of a server farm on the network. The load balancer listens on the needed port, and when a request comes in for the service, the load balancer will forward the request to one of the available servers. When the back-end server receives the request from the load balancer, it will respond to the load balancer. In turn, the load balancer will respond to the client that made the initial request. This makes the load balancer an invisible piece of networking equipment to the client.

Load balancers are aware of the requested protocol and will forward the connection to the correct server. Since not every server will be on equal hardware or be able to handle equal loads, load balancers can be configured with an assigned weight for each server. For example, a single core server can be configured with a lower weight than a quad core server because the quad core server will be able to handle a higher offered load. This ensures that each client will receive the same experience as the next.

Another added benefit of using a load balancer is it helps maintain uptime of the requested services. If one server fails or is taken down for an upgrade, the remaining servers are able to maintain and handle the load until the downed server is online again.

Not all load balancers are created equal. Some have more features than others, but a set of common features does exist between them. As expected, load balancers support all TCP and UDP protocols. A form of SSL acceleration is usually available to improve the speed of encryption. Another common feature is content caching: The load balancer will store objects that are frequently used. When this object is requested, the load balancer will speed up the process by sending the object instead of asking a server for it. Other common features that load balancers support are content compression, session persistence and bandwidth management.

The BreakingPoint Server Load Balancer Test Methodology Journal will demonstrate how to configure a load balancer and set up the BreakingPoint Elite. Get the test methodology at www.breakingpoint.com/resources (less info)