A primary circuit can be fast, SLA-backed and well managed, yet still be unavailable when a duct is damaged, a power fault affects equipment or an upstream network issue occurs. Business backup options should be designed around the real cost of interruption — not treated as an optional add-on at the end of a connectivity project.
Why connectivity resilience matters
For many organisations, internet availability now supports far more than web browsing. It underpins cloud applications, hosted telephony, payment systems, customer service, remote access and communication between sites. When it fails, employees may still be in the office but unable to work effectively.
A suitable backup connection keeps essential services available while the primary circuit is restored.
Why a second connection is not always enough
Two connections do not automatically create resilience. If both services use the same building entry point, street duct, local exchange equipment or provider network, one physical incident can still remove both. Likewise, a backup circuit with insufficient capacity may technically fail over but leave voice calls, critical applications and cloud access unusably slow.
The objective is not simply to have another line. It is to remove credible single points of failure and ensure that the remaining connection can support the services the business needs during an outage. The right design depends on your site, applications, users and recovery priorities.
A single office with twenty cloud-based users has different requirements from a multi-site organisation running centralised systems, IP telephony and customer-facing services. In the latter case, resilience also needs to consider how branches communicate with one another, where security controls sit and whether traffic can be prioritised when capacity is constrained.
A secondary fixed fibre connection
A second fixed connection, such as FTTP or a full fibre leased line, can provide strong performance and predictable capacity. It is often a sensible choice where the business must continue operating near normal levels during an incident, particularly for larger offices, call-heavy teams or sites transferring significant cloud data.
However, resilience depends on route diversity. Ask whether the secondary service enters the premises by a different path and whether it has meaningful separation from the primary circuit beyond the building. Two fibre services from different suppliers may still share infrastructure in the local area. A provider should be able to explain the level of diversity clearly rather than assuming separate bills equal separate routes.
4G or 5G business backup
Mobile connectivity is quick to deploy and physically separate from a wired primary service, making it an effective safeguard against local fibre faults. It is well suited to smaller locations, temporary sites and organisations that need a practical failover option without the lead time or cost of a second fixed circuit.
Its trade-off is variable performance. Signal strength, local congestion, indoor coverage and data allowances all affect service quality. A 5G connection may deliver excellent speeds at one site and be less dependable at another. A proper survey and real-world testing matter more than headline network speeds.
Mobile backup is often best used to support a defined set of critical applications during an outage. This might mean prioritising hosted voice, email, core cloud systems and payment terminals while postponing large downloads, software updates and non-essential guest Wi-Fi traffic. A dedicated mobile broadband service can support this role.
Satellite connectivity
Satellite can provide a useful alternative where terrestrial connectivity is difficult to obtain or where a site needs a genuinely separate access method. It may be relevant for remote locations, construction projects and contingency planning where a wired installation is not viable.
For a typical urban office, it is not usually the first choice. Latency, environmental considerations, installation requirements and service consistency need to be assessed against the applications in use. Real-time voice and certain cloud workloads may be better served by fibre or mobile backup. It remains an option worth considering when geography limits conventional choices.
Diverse carriers and network paths
A resilient design can also use services from different network carriers. This reduces dependency on a single provider's core network and operational processes. Carrier diversity is particularly valuable for critical sites, but it should sit alongside physical route diversity, not replace it.
The strongest arrangements look at the whole service path: the equipment in your comms room, building entry, access network, carrier network, DNS, firewalls and the applications staff rely on. Resilience is only as strong as the component that has not been considered.
Make failover automatic and controlled
A backup connection has little value if someone must notice the outage, reconfigure equipment and ask staff to reconnect manually. Automatic failover should detect a genuine loss of service and move traffic to the secondary path quickly, with clear rules for returning to the primary circuit once it is stable.
SD-WAN can make this more intelligent than a simple on-or-off switch. It can monitor circuit quality, not just whether a connection is present. If packet loss, jitter or latency make hosted calls unreliable, policies can move voice traffic to the healthier path while less sensitive traffic remains where it is. For multi-site organisations, this helps keep branches connected to central systems without treating every application in the same way.
Failover policies should be deliberate. Sending every workload over a mobile backup may exhaust a data allowance or make critical services slow. Prioritising traffic means deciding in advance what must remain available, what can operate at reduced performance and what can pause until the primary service returns.
Size backup capacity around business priorities
The question is not always, "Can the backup match the primary circuit?" In some environments, the answer should be yes. In others, a lower-capacity service is entirely appropriate if it protects the functions that matter most.
Start with the dependency chain. Identify the applications that stop trading, serving customers or communicating with suppliers when connectivity drops. Then consider the number of users who need them, the traffic each application consumes and whether staff can work differently for a short period.
A finance team processing cloud-based transactions, for example, may need reliable access for a small number of users. A contact centre may require enough bandwidth and low enough latency to maintain a large volume of VoIP (voice over IP) calls. A design that is economical for one could be inadequate for the other.
It is also worth accounting for home workers and guest networks. During a primary outage, staff may connect more devices or use services in unexpected ways. Network policies should prevent non-essential activity from displacing operational traffic.
Protect the backup path as carefully as the primary one
Backup connectivity changes the security perimeter. If internet traffic can leave through a secondary circuit, endpoint, identity and email security still need to be supported by consistent network controls. A rushed failover arrangement that bypasses firewall policies or exposes remote access services can introduce avoidable risk at the precise moment the business is under pressure.
The preferred approach is to apply the same security posture across both paths. This includes managed firewall rules, secure remote access, segmentation between business and guest traffic, logging and monitoring. For organisations with several locations, centrally managed policies make it easier to maintain consistency when traffic moves between circuits.
Hosted PBX and VoIP services deserve particular attention. Voice quality relies on stable latency and low jitter, while emergency call handling and number presentation may have specific operational requirements. Test these services on the backup route rather than assuming they will behave identically during an outage.
Test the plan before an outage tests it for you
Business continuity is a process, not a circuit order. Schedule failover tests at sensible intervals and test the tasks people actually need to complete: taking customer calls, accessing cloud files, processing payments, connecting to line-of-business systems and reaching other sites.
Record how long failover takes, whether sessions drop, what happens to voice calls and whether security logs remain visible. Review the results after application changes, office moves, user growth or changes to the primary connection. A backup that was sufficient two years ago may no longer match the way the organisation operates.
The best backup arrangement is the one your team barely notices during an incident, because the priorities, policies and ownership were agreed well before the primary circuit failed.
