When the internet goes down — and it will, eventually — the plan for most businesses is to wait. Staff sit idle, phones stop ringing, card payments fail, cloud applications become unreachable, and someone calls the provider to ask when it'll be back. There's no plan B. There's just a room full of people who can't work.
The cost of that downtime is real. For a 20-person office, even a few hours offline can mean thousands of dollars in lost productivity, missed sales, and frustrated customers. For businesses running cloud phone systems, point-of-sale terminals, or cloud-hosted line-of-business applications, the internet isn't a utility that's nice to have — it's infrastructure that everything else depends on.
Backup internet isn't complicated or expensive. But it does need to be set up properly to actually work when you need it. This article covers the options, the pitfalls, and why path diversity matters more than raw speed.
Why backup internet matters more now than ever
Ten years ago, if the internet went down, you could still make phone calls, process payments on a standalone EFTPOS terminal, and access files stored on a local server. Today, almost none of that is true.
- Phone systems — VoIP and Microsoft Teams Calling run over the internet. No internet means no incoming or outgoing calls. If your phone system is cloud-hosted, it can reroute calls to mobiles during an outage — but only if it knows the connection is down, which requires proper failover.
- Payments — Most EFTPOS and card terminals now connect via IP. No internet, no card payments.
- Cloud applications — Microsoft 365, Google Workspace, Xero, MYOB, CRMs, ERPs — if your team's core tools are in the cloud, they're unreachable without connectivity.
- Remote access — Staff working from home or on the road can't connect to the office VPN or RDP into their desktops.
The question isn't really "can we afford backup internet?" — it's "can we afford to have no plan when the primary goes down?"
Backup internet options — matched to your primary connection
The right backup depends on what your primary connection is. The key principle: your backup should follow a different physical path than your primary, so a single point of failure — a damaged cable, a faulty exchange port, a provider outage — doesn't take out both connections at once.
If your primary is fibre or NBN Enterprise Ethernet
If you're running a dedicated fibre or NBN Enterprise Ethernet service as your primary, a standard NBN Business Internet connection makes an excellent backup. It's a different service class running over different infrastructure (in most cases), and it's affordable enough to keep active permanently rather than only activating it during an outage.
This is the most common setup for Adelaide businesses that need reliable connectivity without enterprise-level costs. The primary fibre handles the workload and carries the SLA. The NBN backup sits ready, and your router fails over automatically if the primary drops.
If your primary is NBN
If your primary connection is already NBN, your backup needs to be a completely different technology — otherwise you're just buying two tickets on the same train. A second NBN service at the same premises typically runs through the same lead-in cable, the same NTD, and the same exchange equipment. If any of those fail, both services go down together.
For NBN-primary businesses, 4G or 5G mobile broadband is the most practical backup. A small business-grade 4G/5G router with an external antenna can be pre-configured and left permanently connected to your network. When the primary drops, the router detects the failure and routes traffic over the mobile network within seconds.
4G backup won't match your NBN for speed or consistency, but it doesn't need to. Its job is to keep critical services running — phones, email, payments, cloud access — until the primary is restored. For most businesses, 20–50 Mbps over 4G is more than enough to keep operating.
For mission-critical connectivity
Some businesses can't afford any downtime at all. Medical practices, financial services, logistics operations, or any business with a private connection to a data centre — these need a higher level of resilience.
The answer is fully diverse fibre circuits — two fibre connections from different carriers, entering the building via different physical paths (ideally different street-level conduits), terminating on different equipment, with automatic failover at the router or SD-WAN layer. If one circuit is cut — whether by a backhoe, a storm, or a provider fault — the other takes over instantly, often without users noticing.
This level of diversity is also essential for businesses running private WAN links between offices, or dedicated circuits to co-location facilities. If your production systems live in a data centre and your office accesses them over a private link, that link is as critical as the systems themselves.
Diverse fibre is more expensive than a 4G backup, but the cost is relative to what's at stake. For a business where an hour of downtime costs more than a month of backup connectivity, the maths is straightforward.
Path diversity — the single most important thing to get right
This is where most backup internet setups fail, and it's the concept that separates a genuine failover solution from an expensive false sense of security.
Path diversity means your primary and backup connections share no common infrastructure. No shared cables, no shared conduit, no shared exchange, no shared upstream provider. If they share anything, that shared element is a single point of failure — and the one thing most likely to take out your backup at the exact moment you need it.
Common diversity failures we see:
- Two NBN services at the same address — They almost certainly share the same lead-in cable from the street to your building. A contractor cuts that cable, and both services go down.
- Two fibre services from different providers, same conduit — Many "diverse" fibre builds enter the building through the same Telstra or NBN pit-and-pipe infrastructure. If that conduit is damaged, both circuits fail. True diversity requires separate physical entry points to the building.
- Same upstream carrier — Two ISPs that both wholesale from the same upstream network may have a common point of failure at the exchange or backhaul level. This is surprisingly common and difficult to verify without help from the carriers themselves.
How failover actually works
Failover is the mechanism that detects the primary connection has failed and switches traffic to the backup. In most business setups, this happens at the router or firewall level.
A dual-WAN router or firewall connects to both your primary and backup internet services. It continuously monitors the primary — either by checking the link state, pinging a known IP address, or monitoring traffic flow. When it detects the primary is down, it reroutes traffic to the backup. When the primary recovers, it switches back.
That's the theory. In practice, there are several things that need to be configured correctly for this to work reliably.
The failover must actually be tested
It's remarkably common for businesses to set up a backup connection, test it once on installation day, and never test it again. Months later, when the primary goes down, they discover the backup SIM has expired, the 4G router firmware has crashed, the backup circuit has been silently disconnected by the provider, or the failover rules on the firewall were overwritten during a firmware update.
Test your failover regularly — at least quarterly. Physically disconnect the primary and confirm everything works on the backup. Then reconnect and confirm it fails back cleanly.
Failback must be graceful
Switching to the backup is only half the problem. Switching back to the primary when it recovers is where things often go wrong.
If the router detects the primary is back and immediately switches all traffic, existing sessions — file transfers, VPN tunnels, VoIP calls, video conferences — can be dropped mid-stream. The better approach is a "preemptive with delay" failback: the router waits for the primary to be stable for a defined period (e.g. 60–120 seconds) before switching back, and ideally routes new sessions to the primary while allowing existing sessions on the backup to complete naturally.
Some routers handle this well out of the box. Others need manual configuration. If your failover was set up by plugging in a second WAN cable and ticking a checkbox, the failback behaviour is worth investigating before you need it.
Not all services will work on the backup
This is the pitfall that catches the most businesses off guard. Your backup connection may have a different public IP address, lower bandwidth, higher latency, or different routing characteristics — and some of your services may not cope.
- Static IP dependencies — If your email server, VPN endpoint, or firewall rules depend on a specific public IP address, switching to a backup connection with a different IP will break those services. Solutions include using dynamic DNS, configuring your VPN to accept connections from multiple IPs, or arranging for your backup to carry a portable IP block.
- VoIP and SIP trunks — SIP trunks are often registered to a specific public IP. When failover changes the IP, registrations can fail and calls stop. Some providers support IP-agnostic registration (credentials-based rather than IP-based), but this needs to be configured in advance, not discovered during an outage.
- Bandwidth-sensitive applications — A 4G backup won't sustain the same load as a 100 Mbps fibre primary. Large file transfers, video conferencing with many participants, and cloud backups may need to be throttled or paused on the backup connection to keep critical traffic flowing.
- Site-to-site VPNs — IPsec tunnels between offices typically bind to specific WAN interfaces. If the tunnel isn't configured to re-establish over the backup WAN, inter-office connectivity will be lost even though both sites have internet access.
The fix for all of these is the same: identify them before an outage, not during one. Walk through every service your business relies on and ask "what happens to this when we're on the backup connection?" Then configure accordingly.
What good looks like
A well-implemented backup internet setup has these characteristics:
- True path diversity — Primary and backup share no common physical infrastructure
- Automatic failover — The switch happens within seconds, without manual intervention
- Graceful failback — When the primary recovers, the transition back is smooth and doesn't drop active sessions
- All critical services tested on the backup — VoIP, VPN, payments, cloud apps have all been verified working over the backup connection
- Regular testing — Failover is tested at least quarterly to confirm everything still works
- Monitoring and alerting — The business (or their IT provider) is notified immediately when failover occurs, so the primary fault can be investigated and resolved rather than ignored
A quick guide to matching your backup to your needs
| Business type | Primary | Recommended backup |
|---|---|---|
| Small office, cloud-first | NBN Business | 4G/5G mobile broadband |
| Medium office, VoIP phones | Enterprise Ethernet | NBN Business or 4G |
| Multi-site, private WAN | Dedicated fibre | Diverse fibre (different path) |
| Data centre connectivity | Private fibre to DC | Fully diverse fibre circuit |
| Retail / hospitality | NBN Business | 4G with EFTPOS priority |
How Caznet can help
We design backup internet solutions as part of every business connectivity engagement — not as an afterthought. Because we supply NBN, Enterprise Ethernet, fibre, and mobile broadband services, we have visibility across the full path from your premises to the internet. That means we can engineer genuine diversity — verifying that your primary and backup follow different physical routes, use different upstream infrastructure, and won't fail together.
We configure failover at the router or firewall level, test it thoroughly, and make sure your phone system, VPN, and critical applications all work correctly on both connections. And when failover does occur, our monitoring picks it up immediately so we can start working on the primary fault before you've even noticed the switch.
If your business doesn't have a backup internet plan — or has one that's never been tested — talk to our team. It's one of the simplest and highest-value improvements you can make to your business continuity posture.