Load Balancing

Load Balancing

What is Load Balancing?

The following 3 are essential within a load balancing scenario:

• End user
• Load balancer
• multiple devices to be load balanced.

Load balancing is a general technique that is used to share or spread work over multiple network resources, devices or components instead of one.

For example, Load balancing can be achieved using multiple (2 or more) computers, servers, ISP/network links, CPUs, hard drives and other resources for optimised use.

Load balancing is performed by specialist hardware or software usually known as a load balancer which is deployed into the network usually between the end user and the network device or component that is utilising the load balancer. Load balancers do share common features however extra features will vary depending on the make and model.

Any incoming requests from end users to the above network devices and components are generally intercepted by the load balancer. The load balancer makes a preconfigured decision as to how, when and where to send the request for the best results.

Load balancing consists of different methods known as algorithms that can be applied to deliver the best solution. Some of the algorithms include the following:

• Round Robin - Each incoming request from individual users is directed to each available network device in turn.
• Weighted Round Robin - If the capacity between the available network devices is different, then the network device with the highest capacity first will receive a higher percentage of requests.
• Least connections - End user requests are forwarded to the network device handling the least percentage of requests.
• Weighted Least connections - End user requests are distributed to the first network device handling the least weight before the other network devices.
• Fastest Response Time - End user requests are passed onto the network device with the fastest response time.
• Adaptive Balancing - The load balancer checks the load upon each network device regularly and if a network device is overloaded, then the end user requests are distributed over all available devices to avoid excessive overloading.

An additional and important common feature of load balancers is failover. With multiple network devices connected to the Load balancer, if a device does fail the load balancer will switch to another available device to ensure business continuity. It is also possible to install two load balancers together and utilise one as the back-up load balancer if the primary load balancer fails - known as high availability.