One of the primary selling points of pfSense is that deploying pfSense routers on our network enhances the overall reliability of the network. A single network component, however – for example, a single router or a single web server – still represents a single point of failure. Even in the absence of hardware failure, a single network component may not be adequate in accommodating the level of traffic on our network. For that reason, we need to consider eliminating single points of failure from our network whenever possible. This process is two-pronged, and involves incorporating both redundancy and high availability:

• Redundancy is defined as the duplication of critical components. Redundancy can be both active and passive. With passive redundancy, we incorporate excess capacity into the network, so that when an individual component fails, resources are still available. An example of this would be having two or more redundant web servers. If one server fails, the website should still be available. Active redundancy involves monitoring components and doing an automatic reconfiguration if a component fails. This might involve, for example, having a spare web server on the network, but not active. When the primary web server goes down, the failure is detected and the spare becomes active. As you may have guessed, both forms of redundancy are implemented in pfSense.
• High availability is defined as ensuring a specified level of operational performance over a prolonged period of time. In practice, it means incorporating some of the same elements as redundancy. Single points of failure are eliminated when possible, and we seek to detect failures when they occur and provide for reliable switch over to the redundant components. Again, we can use pfSense to provide high availability.

pfSense incorporates these functions via load balancing and the Common Address Redundancy Protocol (CARP), which are the main subjects in this chapter.

This chapter will cover the following topics:

• Basic load balancing and CARP concepts
• Server load balancing
• CARP configuration
• An example of both load balancing and CARP
• Troubleshooting