Load balancing can help boost the performance of your website and applications, as it allows more people to use your systems at the same time. Instead of all of the traffic going to a single destination, you can direct traffic to various machines so that no single machine becomes bogged down with handling all of the requests. This can be quite helpful when your systems begin to receive lots of traffic!

What Is Load Balancing?

Load balancing is a method of handling a large amount of traffic to a system (website, web app, database, etc.) in a way that keeps a single server or machine from becoming overloaded, which can greatly slow down your service or even cause it to become unable to handle the traffic at all. Since extremely slow load times and downtime can be especially costly to a business, finding a way to avoid these issues is incredibly important.

Load balancing helps alleviate the traffic congestion by using one or more load balancers that direct the traffic to one of multiple possible machines, all of which have a copy of the content that would otherwise be on a single machine. On a single machine, a large amount of traffic would begin to slow response times pretty quickly. However, using load balancing, each request can be directed to the machine it needs to go to based on your needs, thus keeping the load on any individual machine to a minimum and keeping response times at a maximum.

An example of load balancing using servers (server array). Source: Tractionet.

Deciding How to Direct Requests

With load balancing, you have to decide to which machine a new request is directed. This involves at least a couple of considerations. The two biggest ones are whether or not a machine is responding and the amount of load on the machine.

To determine whether or not a machine is responding, you could use a simple PING, but this isn't necessarily reliable and doesn't offer as many options. Oftentimes, a more advanced service PING, which can determine information such as whether or not a particular service is working on the machine. This allows the load balancer to continue to direct traffic to a machine that may have one service down, but another working. It can simply direct traffic to that machine only if the request is for a working service, and move on to another machine if the request is for a service that is not working.

To determine load different measures can be used — from the current connection count on a machine to the use of the actual response time of the machine to requests. A good load balancing algorithm should take both availability and load into account when deciding how to direct traffic. The actual algorithm will depend on your needs, but getting the right load balancing in place can definitely have a great effect on your response times!

Get PaaS With Morpheus

Looking for a good PaaS service to run your apps and provide you with load balancing options? Morpheus makes provisioning, scaling, and maintenance a breeze. You can provision databases, servers quickly, and have your app up and running in no time! Using the available tools, you can also monitor the various parts of your system to keep track of uptime, response time, and to be alerted if an issue does arise.

The Morpheus interface is clean and easy to use. Source: Morpheus.

Morpheus allows you to provision apps in a single click, and provides ease of use for developers with APIs and a CLI. In addition, backups are also automatic, and you can have redundancy as needed to avoid potentially long waits for disaster recovery to take place.