Failover and redundancy are not quite the same thing
Failover is the process of moving traffic to a backup connection when the primary service fails.
Redundancy is the broader design that removes or reduces single points of failure. It can include diverse internet services, resilient routers and switches, backup power, alternative DNS services, monitoring and documented recovery procedures.
A business can have failover without having meaningful redundancy. Two internet plans connected through the same lead-in cable, router and power supply may still fail together.
The goal is not simply to buy a second service. It is to understand what can interrupt the business and design practical alternatives around the most important risks.
Why internet resilience matters
An internet outage now affects much more than web browsing and email. It can interrupt:
- Cloud phone systems and Microsoft Teams calling
- EFTPOS and online payments
- Cloud applications and file access
- Remote access and VPN connections
- Security camera monitoring and alerts
- Booking, clinical, accounting and customer-management systems
- Communication between offices and remote staff
For a business that relies on cloud services, internet resilience is part of business continuity rather than an optional network feature.
The foundations of a resilient design
A practical internet redundancy design normally includes:
- A reliable primary connection sized for normal business demand.
- A genuinely diverse secondary connection that avoids as many shared failure points as practical.
- A router or firewall that detects faults and fails over automatically.
- Traffic rules that protect essential services when backup capacity is lower.
- Monitoring and alerts so faults do not remain hidden.
- Backup power for the modem, router, firewall, switches and other critical equipment.
- Regular testing to confirm the design still works.
Each component matters. A backup service cannot help if the router has failed, the communications rack has lost power or nobody knows the secondary connection stopped working months earlier.
Choosing diverse internet connections
The best combination depends on availability, budget, performance requirements and the consequences of an outage.
Common Australian designs include:
- NBN primary with 4G or 5G backup — a practical starting point because the services use different access networks.
- Business fibre primary with NBN backup — suitable where the primary service requires stronger performance or service commitments.
- Business fibre primary with 4G or 5G backup — useful where fixed-line and mobile-network diversity are priorities.
- Two fixed-line services — potentially suitable, but only after confirming how much infrastructure they share.
- Dual fibre services with verified path diversity — appropriate for sites where extended downtime would have a serious operational impact.
Different providers do not automatically mean different infrastructure. Two NBN services can share the same building lead-in, local access network or power dependency. Ask what is genuinely diverse rather than relying only on different provider names.
Why 4G and 5G are useful backup options
Mobile internet is often a strong secondary service because it avoids the physical cable used by the primary fixed-line connection.
A well-designed 4G or 5G backup can:
- Be deployed more quickly than another fixed-line service
- Keep cloud phones, EFTPOS and essential applications operating
- Provide enough capacity for reduced but productive business operations
- Avoid a local cable cut affecting both connections
- Remain available while a fixed-line fault is being repaired
Mobile coverage should be tested at the actual site. Signal strength, building materials, antenna placement, congestion and carrier choice all affect performance. An external antenna or a different mobile carrier may be necessary in some locations.
Size backup capacity around essential operations
The backup connection does not always need to match the primary service. It needs enough capacity for the services that must remain usable during an outage.
Start by classifying traffic:
- Critical: voice calls, EFTPOS, core cloud systems, VPN and essential remote access
- Important: email, collaboration, standard SaaS applications and operational browsing
- Deferrable: large backups, software downloads, guest Wi-Fi, streaming and non-essential cloud synchronisation
During failover, the router can prioritise critical applications and restrict or block heavy non-essential traffic. This allows a smaller secondary connection to preserve business operations without attempting to reproduce normal usage completely.
Voice traffic usually requires modest bandwidth, but it is sensitive to packet loss, latency and jitter. Quality and network management matter more than headline download speed.
How automatic failover should work
A suitable router or firewall should actively test whether the primary service is usable. Checking only whether a cable is connected is not enough, because the modem may remain online while access beyond it has failed.
Health checks can use multiple methods, including DNS, ping and HTTP requests. When the primary connection fails the defined tests, the router moves traffic to the secondary service.
A good design should also define:
- How quickly failover occurs
- Which traffic can use the backup connection
- Whether active sessions need to reconnect
- How and when traffic returns to the primary service
- How repeated switching between unstable services is prevented
- Who receives an alert when failover occurs
Most internet sessions will recover, but a live voice call or VPN session may need to reconnect because the public IP address changes. Requirements for seamless session continuity need a more specialised design than standard active-and-standby failover.
Redundancy beyond the internet service
The connection is only one part of the path between a user and a cloud service.
Consider the other possible single points of failure:
Router or firewall
A single failed edge device can take both internet services offline. Higher-availability environments may use redundant firewalls or keep a preconfigured replacement ready.
Network switches and Wi-Fi
Internet failover does not help if the core switch or wireless network has failed. Critical network equipment should be selected, configured and supported as part of the same resilience plan.
Power
A UPS can keep modems, routers, firewalls, switches and phone equipment operating during short power interruptions. Longer outages may require generator support or a defined shutdown plan.
DNS and authentication
Cloud access can appear offline when DNS, identity or authentication services fail. Network design should avoid unnecessary dependencies on one local server or resolver.
Carrier and provider dependencies
Two services may use different access technologies but still share an upstream provider or support process. Perfect independence is rarely practical, but understanding shared dependencies helps the business make an informed decision.
Monitoring and testing
A backup connection that is never monitored or tested should not be assumed to work.
Monitoring should report:
- Primary and secondary connection status
- Failover and restoration events
- Signal quality for mobile services
- Packet loss, latency and jitter
- Data usage and mobile-plan limits
- Router, firewall and UPS health
Schedule periodic failover tests during a controlled window. Confirm that critical services remain usable, alerts are delivered, traffic returns to the primary link correctly and the backup service has not expired, been disconnected or exceeded its allowance.
Document the result and any manual steps. Redundancy is an operational process, not a one-time installation.
Common design mistakes
Buying two services without checking diversity. Different invoices do not guarantee different physical paths or upstream dependencies.
Using a consumer mobile hotspot as the entire continuity plan. It may help in an emergency, but it does not provide automatic failover, central traffic control or dependable monitoring.
Forgetting the router as a single point of failure. Both links can be healthy while one failed edge device leaves the site offline.
Ignoring power protection. A second connection is ineffective when the communications rack loses power.
Allowing all traffic onto a smaller backup service. Cloud backups, guest Wi-Fi and large downloads can overwhelm the link and disrupt calls or payments.
Failing back too quickly. An unstable primary service can cause repeated switching. Sensible timers and health thresholds reduce this behaviour.
Not testing the design. Configuration changes, expired SIMs and failed hardware can silently undermine redundancy.
A practical SMB design
A common design for an Australian office is:
- Business NBN or fibre as the primary service
- A business-grade 4G or 5G service from a diverse network as the secondary
- A router or firewall with automatic health checks and failover
- Traffic prioritisation for cloud voice, EFTPOS and core applications
- Restricted guest access and large downloads during failover
- A UPS supporting the communications rack
- Monitoring with email or mobile alerts
- Scheduled failover testing and documented recovery steps
This design is not completely fault-proof, but it removes several common single points of failure without introducing unnecessary complexity.
When stronger redundancy is justified
Some sites need more than a basic primary-and-backup design. Consider greater resilience where:
- Internet downtime stops revenue or essential customer service
- The site supports critical health, safety or security functions
- Large numbers of staff depend on cloud applications
- Multiple offices rely on the same location or network edge
- Regulatory or contractual obligations require stronger continuity
- Mobile coverage is unsuitable as the only backup path
Options can include verified diverse fibre paths, redundant firewalls, multiple carriers, software-defined WAN, active-active links, generator power and automated application-level testing.
The investment should match the business impact of downtime rather than following a standard package.
In summary
Business internet failover is the automatic switch to another connection. Business internet redundancy is the complete design that makes that switch useful and reduces the chance of one fault taking everything offline.
For many Australian SMBs, a diverse fixed-line primary and 4G or 5G secondary service behind a properly configured router is a practical foundation. Add traffic prioritisation, monitoring, backup power and regular testing, and the business has a far more credible continuity plan than a second connection alone can provide.
Start with the operations that must remain available, identify the likely failure points, and design the simplest level of redundancy that protects them effectively.