A dropped internet service can stop far more than email. EFTPOS terminals may fail, cloud phones can go quiet, staff lose access to business applications, and a regional site may be left without a practical way to trade. Knowing how to configure network redundancy means planning for those failures before they become a costly interruption.
For most Australian businesses, redundancy is not about buying duplicate equipment for every part of the network. It is about identifying what must stay online, then creating an alternative path that activates quickly when the primary service, device or power source fails. The right design depends on your locations, applications, outage tolerance and budget.
Start with the services that cannot go offline
Before selecting a second connection or a new firewall, define what continuity actually means for your business. A medical practice may need cloud-based patient records and phones available at all times. A retailer may prioritise payments and stock systems. A construction business with remote teams may need reliable access to plans, video calls and field reporting.
Map the systems that rely on your network, including internet access, cloud applications, VoIP or cloud PBX, payment terminals, security cameras, remote access and site-to-site connections. For each one, decide how long it can be unavailable. A service that can tolerate a two-hour delay does not need the same investment as a service that must recover in seconds.
It also helps to set two clear targets. Your recovery time objective is how quickly connectivity needs to return. Your recovery point objective is how much data loss is acceptable. Network redundancy mostly addresses recovery time, but the two are closely connected when applications and data are hosted across multiple sites or in the cloud.
How to configure network redundancy in layers
A dependable design rarely relies on one backup. It uses layers, because different failures need different responses. An NBN outage, a damaged local cable, a failed router and a power loss are not the same event.
Use two genuinely independent internet paths
The foundation is a primary and secondary WAN connection. Where possible, choose services that do not share the same physical path, network equipment or upstream carrier. For example, a fixed fibre service paired with 4G or 5G fixed wireless can protect a site when the primary fixed connection is unavailable. In some locations, a second fixed service from a different access network may be suitable. Satellite can also provide an important continuity option for remote and regional operations where other alternatives are limited.
Two connections from the same provider can still be useful, particularly when they are delivered through different technologies. However, asking how each service reaches the premises matters. If both services share the same lead-in cable or street infrastructure, a physical fault can affect both at once.
Your secondary link should be sized for essential operations rather than peak-day convenience. A modest mobile service may keep emails, payments and phones working, but it may not support large cloud backups, high-definition meetings and dozens of staff at the same time. Set expectations around what remains available during failover.
Configure automatic WAN failover
A business-grade router, firewall or SD-WAN appliance should monitor the primary connection and automatically direct traffic to the backup when it detects a fault. Do not rely solely on whether a modem reports an active link. A service can appear connected while it cannot reach the internet or key cloud platforms.
Configure health checks to test reliable external destinations, such as DNS services and major public endpoints. Use more than one destination where the device supports it, so one unreachable website does not trigger an unnecessary failover. Set sensible timers as well. Failover that is too sensitive can switch links during a brief packet-loss event, while failover that is too slow leaves staff waiting through a real outage.
For many small businesses, active-passive failover is the clearest approach: the primary service carries normal traffic and the backup takes over only when required. Larger sites may use active-active connections, where both links carry traffic and capacity is shared. Active-active can improve performance, but it takes more careful policy design and may complicate applications that expect traffic to come from a consistent public IP address.
Protect the router, firewall and local network
A second internet service is of little value if the only firewall fails. For sites where downtime has a high cost, deploy a high-availability firewall pair. The devices share configuration and session information, allowing the standby unit to take over when the active unit has a hardware or software failure.
This setup needs more than two appliances in the same rack. Use separate power supplies where possible, and connect each firewall to redundant switches. Consider the switch layer too. A single unmanaged switch can become the weak point between your internet services and the rest of the business.
Network segmentation remains essential during a redundancy project. Keep staff devices, guest Wi-Fi, voice services, cameras and payment systems separated using VLANs and firewall policies. This limits the impact of a fault or security incident and makes it easier to prioritise critical traffic during a backup-link event.
Prioritise traffic when the backup link is running
When a backup service has less bandwidth than the primary connection, quality of service policies become practical rather than optional. Give priority to voice, payments, remote access, DNS and the cloud applications your team needs to operate. Restrict or deprioritise software updates, large downloads, personal streaming and cloud backup jobs until the main service returns.
For VoIP, check that your firewall is configured to handle voice traffic correctly and that your phone system has a continuity plan. Cloud PBX services can often route calls to mobiles or another site during a prolonged outage. That fallback should be documented and tested with reception and key team members, not just configured in an admin portal.
If your team uses a static IP address for remote access, whitelisting or site-to-site VPNs, consider what happens when traffic moves to the secondary link. You may need a managed SD-WAN solution, dynamic DNS, a cloud security service or a secondary static IP arrangement to preserve access. This is a common point where basic failover works for web browsing but not for business-critical services.
Do not overlook power and physical risks
Internet redundancy will not help if the network cabinet has no power. Install an uninterruptible power supply for the modem or network termination device, router or firewall, switches and Wi-Fi equipment. Calculate runtime based on the actual equipment load. A small UPS may bridge a short interruption, while a longer outage may require generator support or a planned shutdown process.
Physical placement matters as well. Keep primary and backup equipment protected from water, heat, accidental disconnection and unauthorised access. For multi-site businesses, avoid placing all network equipment in one comms cupboard if another practical location can reduce a single point of failure.
Test failure conditions, not just the installation
A redundancy setup is only proven when it has been tested under controlled conditions. Schedule a failover test at least twice a year and whenever you make major changes to your internet service, firewall, cloud phone system or network topology.
During the test, disconnect the primary WAN service and measure how long it takes for key functions to recover. Make a payment test, place and receive calls, connect through the VPN, access core cloud applications and check remote monitoring. Then restore the primary service and confirm that failback occurs as intended. Some organisations prefer manual failback to avoid moving traffic back to a connection that is still unstable.
Record the results, including any applications that fail or behave differently. A short runbook should state who receives outage alerts, who contacts the provider, how staff are informed and which services are restricted while operating on backup. Clear ownership prevents a minor incident from becoming a long period of uncertainty.
Avoid the common redundancy traps
The most frequent mistake is assuming a second service automatically means continuity. If the secondary link has no automatic failover, no tested configuration or insufficient capacity, it is a spare connection rather than a continuity solution.
Another issue is creating hidden single points of failure: one power board, one firewall, one internal switch, one lead-in cable or one person who knows the configuration. The goal is not perfection. It is to reduce the most likely and most damaging points of failure within a sensible budget.
Security should also remain in place during failover. The backup connection needs the same firewall rules, web filtering, VPN access controls and monitoring standards as the primary service. A rushed mobile hotspot may restore access, but it can also bypass the policies that protect your business.
For organisations with multiple sites, SD-WAN can simplify this work by applying traffic policies, link monitoring and security controls consistently across each location. InfiNET Broadband can help businesses assess available access technologies and build continuity arrangements that match the way they operate, from a single office to a distributed network.
The best time to test an outage is when you can choose the moment. Build redundancy around the services your people and customers depend on, test it regularly, and your business will be far better placed to keep moving when the unexpected happens.