Business continuity guide
Internet Failover Guide
Modern businesses depend on cloud applications, payments, voice, bookings, remote access and communications. A second connection can reduce the impact of an ISP outage—but only when the backup route, gateway behaviour and critical applications have been designed and tested together.
Treat the complete network as one system.
- Identify the services that must remain available
- Choose meaningful provider or technology diversity
- Automate health checks and failover
- Test the backup under realistic conditions
Two routes. One managed decision point.
The gateway checks connection health and moves essential traffic when the primary route fails.
A backup is useful only when its route, capacity and critical applications have been tested together.
A practical approach
Evidence first. Changes second.
Why internet connections fail
Outages may be caused by fibre damage during roadworks, failed provider equipment, local power loss, routing or exchange faults, planned maintenance, configuration errors or wider network disruption. A circuit with a strong service history can still be affected by events outside the building.
Define what must keep working
List the operational systems that cannot wait for the primary service to return. These may include EPOS and card payments, cloud applications, VoIP, booking systems, VPN access, alarms or remote CCTV. The backup design and bandwidth should reflect these priorities rather than trying to recreate every non-essential activity.
Diversity matters
Two services are not necessarily independent. They may share ducts, exchanges, wholesale infrastructure or power. Where availability matters, consider a different provider, physical route or access technology such as managed 5G, and document the limitations of the available options.
Automatic failover needs application awareness
A managed gateway can test connection health and move eligible traffic to the secondary route. Static IP addresses, inbound services, VPNs, voice and provider restrictions may require specific design. The network should also return to the primary circuit in a controlled way.
Test, monitor and revisit
A backup that has never been tested is an assumption. Run controlled failover tests, confirm critical applications, monitor both links and check mobile signal or backup capacity over time. Update the design when applications or site requirements change.
Use this on site
Failover planning questions
A concise starting point for a useful internal review or a conversation with a network specialist.
- 01Which systems are business-critical?
- 02How much download and upload capacity do they require?
- 03Are the primary and backup routes genuinely diverse?
- 04Do VPNs, voice or static IP services need special handling?
- 05How will users know the site is on backup?
- 06When was the last controlled failover test?
Put the guidance into practice
Relevant service and real-world example.
Common questions
Useful answers before you act.
Does internet failover happen automatically?
It can. A correctly configured gateway monitors connection health and moves eligible traffic when the primary circuit fails. The exact behaviour depends on the design and applications involved.
Is 5G suitable as a backup connection?
It can be an effective diverse route where coverage, capacity and provider conditions are suitable. Signal quality and performance should be assessed at the site rather than assumed.
Does the backup need the same speed as the primary circuit?
Not always. It needs enough capacity for the services that must continue during an outage, with sensible headroom. Non-essential traffic can be restricted while failover is active.