Many businesses assume that installing a second internet connection gives them a redundant network. That is not always the case.
Two services may have different provider names while still relying on the same wholesale carrier, fibre route, building entrance, local network facility or electrical supply. A single damaged cable, equipment failure or utility outage could therefore interrupt both connections at the same time.
Backup internet becomes genuinely resilient when the complete path between the business and the internet has been assessed. That includes the carriers, access technologies, physical routes, network equipment, power, addressing and operating procedures behind both services.
In practical terms: Backup internet is truly redundant when a single predictable failure is unlikely to disable both the primary and secondary connections.
Backup Internet and Internet Redundancy Are Not Identical
A backup connection is any secondary service that can be used when the primary connection is unavailable. Internet redundancy goes further. It considers whether the secondary service is sufficiently independent from the primary service to survive the same outage.
For example, a business might purchase one connection from its primary cable provider and another from a reseller. The invoices and support contacts may be different, but both services could still use the same local cable infrastructure.
The same problem can occur with fibre. Two providers may use the same street conduit, building entrance, wholesale network or upstream carrier. A construction cut or equipment failure could affect both services.
Businesses should therefore evaluate a backup connection in relation to their existing business internet service, rather than evaluating the second connection in isolation.
1. Carrier and Network Independence
The primary and backup connections should be assessed at both the retail and wholesale levels.
The company selling the service is not always the company operating the underlying network. Internet providers, resellers and managed-service companies may all deliver services over infrastructure owned by the same wholesale carrier.
A resilient design should identify:
- The retail internet service provider
- The underlying wholesale carrier
- The local access-network operator
- The upstream internet networks
- The autonomous system numbers, or ASNs, involved
- The cellular carrier used for an LTE or 5G backup service
Where practical, the primary and backup services should avoid dependence on the same wholesale carrier, upstream network or cellular operator.
This reduces the risk that a provider-wide outage, routing problem or shared network failure interrupts both services.
2. Last-Mile and Physical-Path Diversity
Carrier diversity is useful, but it does not necessarily create physical diversity.
Two carriers may use different networks beyond the local area while still entering the building through the same conduit, utility pole, riser or communications room. Damage to that shared infrastructure could disconnect both services before their paths ever separate.
A physical-path assessment should consider:
- The last-mile medium, such as fibre, cable, copper or wireless
- The street-level route used by each connection
- Underground conduits and utility ducts
- Overhead poles and aerial cable routes
- Building entrances and demarcation points
- Internal risers and communications rooms
- Local headends, central offices and aggregation equipment
Combining different access technologies can improve physical separation. Common examples include:
- Cable internet with LTE or 5G backup
- Fibre internet with fixed wireless or microwave backup
- Wired business internet with LEO satellite backup
- Shared fibre with a separately routed dedicated fibre service
Different technologies do not guarantee independence, but they often reduce the amount of shared last-mile infrastructure.
Physical diversity should be investigated and documented rather than assumed.
3. Equipment and Power Resilience
Two functioning internet circuits provide little protection when they both depend on the same failed router, firewall, network switch or power source.
The equipment that detects an outage and switches traffic to the secondary connection must remain operational during the failure.
A complete backup internet design should account for:
- Router and firewall availability
- Power supplies and electrical circuits
- Uninterruptible power supply capacity
- UPS runtime under the expected equipment load
- Generator availability and startup time
- Network-switch and wireless-access-point power
- Cellular modem and antenna placement
- DNS availability and configuration
A cellular backup service, for example, will not help during a power outage if the modem, firewall, switches and access points lose power immediately.
The business should determine how long essential network equipment must remain available and size its UPS or generator strategy accordingly.
4. Public IP Addresses and Application Continuity
Not every failover service behaves the same way when the primary connection fails.
Basic failover solutions may move traffic to a backup service with a different public IP address. New connections can often be established, but active sessions may be interrupted.
This may affect:
- VoIP calls
- Site-to-site and remote-access VPNs
- Payment terminals and point-of-sale systems
- Cloud applications
- Remote desktop sessions
- Hosted services
- IP-allowlisted applications
- Inbound connections
A continuity plan should define:
- Whether the public IP address changes during failover
- Whether active sessions remain established
- Whether inbound services remain reachable
- How quickly an outage is detected
- How quickly traffic moves to the secondary connection
- How traffic returns when the primary service is restored
- Whether failback causes a second interruption
- How firewall, VPN and DNS rules behave during the transition
A company that primarily uses email and web applications may tolerate a short reconnection period. A healthcare clinic, call centre, retailer or cloud-dependent business may require static-IP continuity and preservation of active sessions.
Fidalia’s managed backup internet services include different levels of failover depending on the required switching time, public-IP behaviour and application-continuity requirements.
5. Available Bandwidth During an Outage
A backup connection does not always need to match the full capacity of the primary service. It must, however, provide enough bandwidth for the applications and users that need to remain operational.
Capacity planning should identify:
- Critical users and departments
- Minimum download and upload requirements
- VoIP call volume
- Video-conferencing requirements
- VPN traffic
- Point-of-sale and payment traffic
- Cloud applications and hosted systems
- Guest Wi-Fi and non-essential traffic
- Cellular data allowances and potential overage charges
Quality-of-service and traffic-control policies can reserve backup capacity for essential systems while restricting less important uses.
During an outage, a business might temporarily limit:
- Guest Wi-Fi
- Software and operating-system updates
- Cloud backups
- Large file transfers
- Streaming services
- Non-essential video traffic
The appropriate backup capacity depends on what the business expects to keep running, not simply the speed of the primary connection.
6. Monitoring, Testing and Reporting
A secondary connection may appear healthy while sitting idle, but that does not prove that the complete failover process will work when the primary service fails.
Problems can develop in firewall rules, routing, cellular data plans, DNS, public-IP assignments, cabling and power systems without being noticed.
A managed continuity plan should monitor and document:
- Primary-connection availability
- Backup-connection availability
- Latency, packet loss and connection quality
- Outage-detection time
- Failover time
- Public-IP continuity
- Application performance during failover
- Failback behaviour
- Test frequency
- Test results
- Failures and remediation actions
Testing should include the transition to the secondary service and the return to the primary service. A system may fail over successfully but cause an interruption when the primary circuit is restored.
Regular testing turns backup internet from an unverified assumption into a documented business-continuity control.
A Backup Internet Redundancy Checklist
| Area | Questions to Confirm |
|---|---|
| Internet providers | Are the primary and backup services sold and operated by different providers? |
| Wholesale carriers | Do both services depend on the same underlying carrier? |
| Last-mile media | Do the connections use different access technologies? |
| Physical routes | Do the services use different conduits, poles, entrances or risers? |
| Upstream networks | Do both connections depend on the same upstream network or ASN? |
| Local facilities | Do the services share a headend, central office or aggregation point? |
| Network equipment | Could one router, firewall or switch failure disable both services? |
| Power | Will the backup connection and local network remain powered during an outage? |
| Public IP | Does the public IP change, and will that affect applications or inbound access? |
| Failover | How quickly is failure detected, and how quickly does traffic switch? |
| Failback | What happens when the primary service returns? |
| Capacity | Can the backup service support essential users and applications? |
| Monitoring | Are both connections continuously monitored? |
| Testing | How often is failover tested, documented and remediated? |
How Fidalia Approaches Backup Internet Design
A backup internet design should begin with the business’s operational requirements, not with a specific connection type.
Fidalia evaluates:
- The existing primary internet service and underlying network
- The applications that must remain available
- The acceptable failover and recovery time
- Public-IP and session-continuity requirements
- Available carriers and access technologies at the address
- Potential physical and carrier diversity
- Required bandwidth during an outage
- Power, equipment, monitoring and testing requirements
The result may use LTE or 5G, cable, DSL, fibre, fixed wireless, microwave or satellite. The right option depends on the primary connection, the location and the consequences of downtime.
Review Fidalia’s business internet options to understand the available access technologies, or explore OnePort backup internet for managed failover, static-IP continuity and sub-second switching options.
Frequently Asked Questions
Is a second internet connection automatically redundant?
No. A second connection may still depend on the same wholesale carrier, physical route, building entrance, local network equipment or power source as the primary service. Those shared dependencies should be identified before the connection is considered redundant.
Should backup internet use a different provider?
Using a different provider can improve resilience, but the provider name alone is not enough. The underlying wholesale carrier, last-mile network, physical path and upstream infrastructure should also be assessed.
What is internet path diversity?
Internet path diversity means that the primary and backup connections use sufficiently separate physical and network routes. This may include different access media, street routes, building entrances, carriers, headends and upstream networks.
Will backup internet work during a power outage?
Backup internet will only work during a power outage if the modem, router, firewall, switches and other required equipment remain powered. A UPS or generator may therefore be necessary in addition to the secondary internet connection.
Will VoIP calls and VPN sessions remain connected during failover?
It depends on the failover design. Basic solutions may change the public IP address and interrupt active sessions. Static-IP or session-persistent failover can preserve more applications, but the exact behaviour should be tested for each environment.
How much bandwidth does a backup internet connection need?
The backup connection should have enough download and upload capacity for essential users and applications. It does not always need to match the primary service when non-essential traffic can be restricted during an outage.
How often should internet failover be tested?
Failover should be tested on a defined recurring schedule and after material changes to the network. Testing should verify outage detection, switching time, application performance, public-IP behaviour and failback to the primary service.