How to Set Up Network Redundancy for Business Continuity

Every business depends on its network. Email, cloud applications, VoIP phone systems, payment processing, remote access — all of it flows over your internet connection and internal infrastructure. When that network goes down, your business effectively ceases to operate.

Yet a surprising number of UK businesses — particularly SMEs — run with a single point of failure at virtually every layer of their network stack. One internet connection from one provider. One router. One core switch. One path between the comms room and the office floor. A single failure at any point takes the entire organisation offline.

Network redundancy is the practice of building backup paths, devices, and connections into your infrastructure so that when something fails — and it will — your business continues operating without interruption. It’s not about preventing failure entirely; it’s about making failure invisible to your users and clients.

This guide covers everything a UK business needs to know about setting up network redundancy for genuine business continuity. We’ll walk through the three core types of redundancy, explain dual ISP configurations, BGP multihoming, 4G/5G failover, redundant switches and routers, first-hop redundancy protocols like HSRP and VRRP, and link aggregation. We’ll also cover how to properly test your failover systems and provide a realistic cost-versus-risk analysis in pounds sterling to help you make informed decisions.

Understanding the Three Types of Network Redundancy

Before diving into specific technologies, it’s important to understand that network redundancy operates at three distinct levels. A truly resilient network addresses all three — because redundancy at one layer is undermined if another layer has a single point of failure.

1. Link Redundancy

Link redundancy means having multiple network connections available so that if one link fails, traffic automatically shifts to another. This is the most commonly understood form of redundancy and the one most businesses address first.

At its simplest, link redundancy means having two internet connections from different providers. But it also applies internally — dual uplinks from access switches to your core switch, multiple fibre paths between buildings, or redundant connections between your network and cloud services.

The key principle is that redundant links must follow physically separate paths. Two broadband connections from different providers that both enter your building through the same BT Openreach duct aren’t truly redundant — a single digger hitting that duct takes both connections out simultaneously. Ideally, your second connection should enter the building via a different route entirely: a different street, a different riser, or a completely different technology such as cellular.

2. Device Redundancy

Device redundancy means having duplicate network hardware so that if a router, switch, or firewall fails, an identical standby device takes over. This includes redundant power supplies within devices, stacked switch configurations, and hot-standby appliances.

Device failures are more common than many businesses realise. Power supplies blow, fans seize, firmware bugs cause crashes, and hardware simply wears out over time. Without device redundancy, a £300 switch failure can take down a £3 million business for hours — sometimes days, if replacement hardware isn’t readily available.

3. Path Redundancy

Path redundancy means having multiple routes for data to travel between any two points on your network. Even if you have redundant links and redundant devices, a single path between them means a cable fault or port failure creates an outage.

Path redundancy is achieved through technologies like Spanning Tree Protocol (STP), Equal-Cost Multi-Path routing (ECMP), and SD-WAN overlay networks. These create multiple available routes and automatically reroute traffic when a path becomes unavailable. In well-designed networks, path redundancy is invisible to users — traffic reroutes in milliseconds without any perceptible disruption.

Dual ISP Setups: Your First Line of Defence

For most UK businesses, the single biggest point of failure is their internet connection. One leased line, one broadband circuit, one provider — and when it goes down, everything stops. A dual ISP setup eliminates this vulnerability by maintaining two independent internet connections from different service providers.

Active–Passive Configuration

In an active–passive setup, your primary connection handles all traffic under normal conditions. The secondary connection sits idle, monitoring the primary link’s health through periodic health checks. When the primary fails, the secondary activates and takes over all traffic. When the primary recovers, traffic shifts back automatically.

This is the simpler and more affordable approach. It works well for businesses that have a fast primary connection (such as a leased line) and a more modest backup (such as FTTP broadband or 4G/5G). The downside is that your secondary connection provides no value during normal operation — you’re paying for capacity that sits unused most of the time. Despite this, many businesses find the peace of mind well worth the monthly outlay.

Active–Active Configuration

In an active–active setup, both connections carry traffic simultaneously. Your firewall or SD-WAN appliance distributes traffic across both links based on policies you define — for example, routing VoIP traffic over the leased line for guaranteed quality, and directing general web browsing over the broadband connection. If either link fails, the surviving connection absorbs all traffic without interruption.

Active–active is the more cost-effective approach because you get value from both connections every day, not just during outages. It also provides greater aggregate bandwidth under normal conditions. However, it requires more sophisticated configuration and a firewall or router capable of policy-based routing and session persistence.

Choosing Your ISPs

The most critical factor in a dual ISP setup is genuine infrastructure diversity. In the UK, the majority of business broadband and many leased line services run over BT Openreach infrastructure, regardless of which ISP you purchase from. If both your connections rely on Openreach, an exchange failure or duct damage could take both down simultaneously.

For true diversity, consider pairing an Openreach-based service with a provider using independent infrastructure — such as Virgin Media (which operates its own cable network), CityFibre, Hyperoptic, or a wireless provider. Alternatively, pair a wired connection with 4G/5G cellular, which uses entirely separate mobile network infrastructure.

BGP Multihoming: Enterprise-Grade Internet Redundancy

For businesses that host public-facing services, run e-commerce platforms, or simply cannot tolerate any internet downtime, BGP (Border Gateway Protocol) multihoming represents the gold standard for internet redundancy.

BGP multihoming involves obtaining your own Provider Independent (PI) IP address space from RIPE NCC (the European internet registry) and your own Autonomous System Number (ASN). Your IP addresses are then announced to the internet via two or more upstream ISPs simultaneously. If one ISP fails, the internet automatically routes traffic to you via the other — with no DNS changes, no IP address changes, and no disruption to active sessions.

The advantage of BGP multihoming is its seamlessness. Your public IP addresses don’t change when a provider fails, which means DNS records, SSL certificates, VPN endpoints, and firewall rules all continue working without modification. For businesses hosting websites, APIs, or customer-facing applications, this is transformative — your services remain reachable even during a complete ISP outage.

Requirements and Costs

BGP multihoming comes with significant requirements. You need a minimum /24 subnet (256 IP addresses) to announce routes on the public internet, which costs approximately £1,500–£2,500 per year from RIPE NCC. You also need BGP-capable routers at your premises and ISPs willing to peer with your ASN — not all UK providers offer this service. Typical setup costs range from £3,000 to £8,000, with ongoing costs of £200–£500 per month on top of your dual ISP subscriptions. This makes BGP multihoming most appropriate for businesses with annual revenue above £2–3 million, where the cost of downtime easily justifies the investment.

4G/5G Failover: The Essential Safety Net

Cellular failover using 4G LTE or 5G is arguably the most cost-effective form of internet redundancy available to UK businesses today. It’s quick to deploy, requires minimal infrastructure changes, and provides a genuinely independent backup path that’s completely unaffected by fibre cuts, exchange failures, or ISP outages.

How It Works

A 4G/5G failover solution consists of a cellular router or modem connected to your firewall or primary router. Under normal conditions, all traffic flows over your primary wired connection. The failover device continuously monitors the health of the primary link — typically using ICMP pings to known-reliable external hosts. When the primary connection fails, the failover device activates its cellular connection and your router switches traffic to the backup path.

Modern enterprise 4G/5G routers from manufacturers like Cradlepoint, Draytek, and Peplink can achieve failover times of 15–45 seconds. Some SD-WAN solutions reduce this to under 5 seconds by maintaining a persistent “warm” cellular tunnel alongside the primary connection, ready to carry traffic instantly.

Choosing the Right Cellular Provider

The critical consideration is network diversity. If your primary ISP uses BT Openreach infrastructure (as most UK broadband and leased line providers do), your 4G/5G backup should use a mobile network that doesn’t share that infrastructure. Three, Vodafone, and O2 operate fully independent mobile networks and make excellent backup choices.

For businesses in central London and other major UK cities, 5G provides speeds that rival many wired connections — often 200–500Mbps download. Outside urban areas, 4G LTE typically delivers 30–80Mbps, which is more than adequate for most business operations during a temporary failover scenario. Budget approximately £40–£100 per month for a business-grade cellular data plan with sufficient allowance for failover use.

Limitations to Consider

Cellular failover has higher latency than wired connections (typically 20–50ms versus 5–15ms for a leased line), which can affect latency-sensitive applications like VoIP and video conferencing. Data allowances, while generous, are not truly unlimited on most business plans — sustained heavy use during a prolonged primary outage could exhaust your cellular data cap. And cellular signal strength can be unreliable in certain buildings, particularly those with metal cladding, thick concrete walls, or below-ground offices. Always test signal strength at your specific premises before committing to a cellular backup solution, and consider an external antenna if indoor reception is marginal.

Redundant Switches and Routers

Your internal network hardware represents another critical layer of potential failure. A single core switch handling all internal traffic is a significant single point of failure — if it dies, every user on the network loses connectivity even if your internet connection is perfectly healthy.

Switch Stacking

Switch stacking connects two or more physical switches so they operate as a single logical unit. If one switch in the stack fails, the remaining switches continue operating and the network stays up. Stacking also provides greater port density and simplified management — you configure and manage one logical device rather than multiple independent switches.

Most enterprise switch manufacturers offer stacking capabilities. Cisco Catalyst, HPE Aruba, Juniper EX, and Meraki MS all support stacking with varying levels of sophistication. For a typical UK SME, a two-switch stack at the core provides excellent redundancy at a reasonable cost — typically £2,000–£6,000 for a pair of managed Layer 3 switches, depending on port count and throughput requirements.

Router and Firewall Redundancy

Your edge router or firewall — the device connecting your internal network to the internet — is another critical single point of failure. For businesses running a single router/firewall appliance, a hardware failure means total loss of internet connectivity regardless of how many ISP connections you have.

The solution is deploying a pair of routers or firewalls in a high-availability (HA) cluster. Vendors like Fortinet, SonicWall, Sophos, and pfSense all support HA configurations where a secondary unit mirrors the primary’s configuration and takes over seamlessly if the primary fails. Session state is typically synchronised between the pair, meaning active VPN tunnels, firewall sessions, and NAT translations survive the failover without dropping.

Power Supply Redundancy

Don’t overlook power. Enterprise-grade switches and routers often support redundant power supplies — two independent PSUs in a single chassis, each capable of powering the device alone. If one PSU fails, the other keeps the device running without interruption. Pair this with a UPS (Uninterruptible Power Supply) to protect against mains power failures, and you’ve eliminated two of the most common causes of unexpected hardware outages. A properly sized UPS for a small network comms cabinet typically costs £500–£1,500 and provides 15–30 minutes of runtime — enough to ride out brief power cuts or shut down gracefully during extended outages.

HSRP and VRRP: First-Hop Redundancy Protocols

When you deploy two redundant routers or firewalls, your network devices need a way to know which one to use as their gateway — and how to switch between them transparently. This is where first-hop redundancy protocols come in.

How They Work

Both HSRP (Hot Standby Router Protocol) and VRRP (Virtual Router Redundancy Protocol) create a virtual IP address that is shared between two physical routers. Your network devices — PCs, laptops, phones, printers — are configured to use this virtual IP as their default gateway. Behind the scenes, one router is designated as the active device and handles all traffic, while the other monitors the active router’s health via periodic heartbeat messages.

If the active router fails, the standby router detects the failure (typically within 3–10 seconds, configurable down to sub-second intervals) and assumes the virtual IP address. From the perspective of your network devices, absolutely nothing has changed — their default gateway address is identical. Traffic simply flows through the newly active router without any user intervention, IP reconfiguration, or noticeable disruption.

HSRP vs VRRP: Which to Choose

HSRP is Cisco’s proprietary protocol, available only on Cisco equipment. VRRP is an open standard (defined in RFC 5798) supported by virtually all enterprise router and switch vendors including Juniper, HPE Aruba, Fortinet, and MikroTik. Functionally, they achieve the same outcome with very similar performance characteristics. The choice typically depends on your existing hardware — if you’re a Cisco-only environment, HSRP integrates naturally and offers a few Cisco-specific enhancements; if you run mixed-vendor or non-Cisco equipment, VRRP is the clear choice for interoperability.

Link Aggregation (LAG / LACP)

Link Aggregation, implemented through the Link Aggregation Control Protocol (LACP, IEEE 802.3ad), bundles multiple physical network connections into a single logical link. This provides both increased bandwidth and link-level redundancy in a single, elegant solution.

For example, bonding four 1Gbps connections between your core switch and a critical server creates a single 4Gbps logical link. If one of the four physical cables or ports fails, the aggregate link continues operating at 3Gbps — reduced capacity, certainly, but critically no outage. Users experience a slight performance reduction rather than a complete loss of access.

Where to Deploy Link Aggregation

Link aggregation is most valuable at network bottleneck points:

• Core-to-distribution switch uplinks: Bonding multiple connections between your core and floor switches prevents a single uplink failure from isolating an entire floor or department
• Server connections: Critical servers benefit enormously from bonded NICs, providing both hardware redundancy and increased throughput for file serving, databases, and applications
• Storage network links: NAS devices and SAN storage frequently support LACP for higher aggregate bandwidth and fault tolerance on storage-intensive workloads
• Inter-site connections: If you operate multiple offices connected by dark fibre or MPLS circuits, LAG provides resilience and capacity on those critical inter-site links

LACP is supported by virtually all managed switches and most enterprise-grade servers, NAS devices, and firewalls. Configuration is straightforward — both ends of the aggregated link must be configured to use LACP, and the protocol handles the negotiation, member detection, and load distribution automatically. The cost is essentially just the additional cables and switch ports, making link aggregation one of the most affordable redundancy measures available to any business.

Testing Your Failover Systems

Redundancy that hasn’t been tested isn’t redundancy — it’s a hope. We’ve encountered countless businesses that invested in dual ISP connections, redundant firewalls, and backup cellular links, only to discover during an actual outage that the failover didn’t work as expected. A configuration change had broken the health monitoring. A firmware update had reset the failover policy. The cellular SIM had expired because nobody remembered to renew it. A cable had been accidentally unplugged during a routine office clean.

What to Test and How Often

Every redundant component in your network should be tested on a regular, documented schedule. The table below outlines recommended testing procedures and frequencies for common redundancy measures:

Component	Test Method	Frequency
Dual ISP failover	Disconnect primary link; verify traffic shifts to secondary within expected time	Monthly
4G/5G cellular backup	Disable all wired connections; confirm cellular activates and carries traffic	Monthly
HSRP/VRRP failover	Power off active router; verify standby assumes virtual IP within threshold	Quarterly
Switch stack resilience	Remove one switch from the stack; confirm network continuity for all users	Quarterly
Firewall HA cluster	Fail the primary firewall; verify secondary takes over with sessions intact	Quarterly
Link aggregation (LACP)	Disconnect one cable in the LAG bundle; verify traffic continues on remaining links	Bi-annually
UPS battery backup	Simulate mains power failure; verify UPS sustains all critical network equipment	Bi-annually
BGP route failover	Withdraw routes on primary ISP; verify traffic reroutes via secondary provider	Quarterly

Document Everything

Every failover test should be formally documented with the date, what was tested, the observed failover time, whether the failover was successful, and any issues discovered during the process. This documentation serves multiple critical purposes: it demonstrates due diligence for compliance and insurance audits, it creates a performance baseline for detecting degradation over time, and it provides invaluable troubleshooting context when real failures eventually occur. Store test results alongside your network documentation and review them during quarterly IT reviews.

Cost vs Risk Analysis: What Network Redundancy Actually Costs

One of the biggest barriers to implementing network redundancy is the perceived cost. Many UK business owners assume it requires enterprise-level budgets and complex infrastructure. But when you compare the investment against the real cost of downtime, the mathematics almost always favours redundancy — often dramatically so.

The following table shows realistic UK costs for common redundancy measures, alongside the downtime risk each one mitigates. All figures are based on current 2026 market prices for a typical 25–50 person UK business.

Redundancy Measure	Setup Cost	Monthly Cost	Annual Total	Downtime Prevented
Second broadband (FTTP)	£0 – £100	£40 – £70	£480 – £940	8 – 24 hrs/yr
4G/5G cellular failover	£200 – £500	£40 – £100	£680 – £1,700	8 – 24 hrs/yr
Dedicated leased line (secondary)	£500 – £2,000	£200 – £500	£2,900 – £8,000	12 – 48 hrs/yr
HA firewall pair	£2,000 – £8,000	£0 – £300	£2,000 – £11,600	4 – 16 hrs/yr
Core switch stack (pair)	£2,000 – £6,000	£0	£2,000 – £6,000	4 – 12 hrs/yr
SD-WAN overlay	£500 – £2,000	£100 – £400	£1,700 – £6,800	12 – 36 hrs/yr
BGP multihoming (full)	£3,000 – £8,000	£200 – £500	£5,400 – £14,000	24 – 72 hrs/yr
UPS + redundant PSUs	£500 – £2,000	£0	£500 – £2,000	2 – 8 hrs/yr

The ROI Calculation

Consider a 30-person professional services firm in London. Their hourly downtime cost is approximately £4,500 (based on average revenue per employee plus productivity loss and reputational impact). Over the past year, they experienced three internet outages totalling 14 hours, two switch-related issues causing 4 hours of disruption, and one firewall failure that took 6 hours to resolve. Total downtime: 24 hours. Total estimated cost: approximately £108,000.

Implementing a comprehensive redundancy solution — dual ISP with 4G failover, an HA firewall cluster, and stacked core switches — would cost approximately £8,000–£15,000 in the first year and £3,000–£6,000 annually thereafter. Against £108,000 in annual downtime costs, the return on investment is between 700% and 1,350% in the first year alone.

Even for smaller businesses with lower per-hour downtime costs, the equation almost always favours investment in redundancy. A basic dual ISP setup with 4G failover can be implemented for under £1,500 in the first year — which is less than a single hour of downtime costs for most 10-person businesses. When you frame redundancy as insurance against a near-certain event (network failure), rather than an optional luxury, the investment case becomes overwhelming.

Building Your Redundancy Strategy: A Phased Approach

You don’t need to implement everything simultaneously. A phased approach allows you to address the highest-risk areas first while spreading costs over a manageable timeline.

Phase 1 — Internet Redundancy (Month 1): Add a second internet connection from an independent provider, configure automatic failover on your existing firewall, and deploy a 4G/5G cellular backup. This single phase eliminates the most common cause of business network outages and delivers the highest ROI of any redundancy investment.

Phase 2 — Device Redundancy (Months 2–3): Deploy an HA firewall pair and stack your core switches. Add redundant power supplies where your hardware supports them. Install UPS units for all critical network equipment in your comms cabinet or server room.

Phase 3 — Path Redundancy (Months 3–6): Implement LACP on critical uplinks between switches and servers. Configure VRRP or HSRP for gateway redundancy if you have multiple routers. Review your internal cabling infrastructure for hidden single points of failure. Consider SD-WAN for intelligent, policy-driven traffic management across your redundant links.

Phase 4 — Advanced Redundancy (Months 6–12): If your business requirements justify the investment, explore BGP multihoming for seamless internet failover and consider geographic redundancy across multiple physical sites for disaster recovery scenarios.

At each phase, test thoroughly, document your configurations meticulously, and establish a regular ongoing testing schedule. Network redundancy is not a one-time project that you complete and forget — it’s an ongoing operational commitment to maintaining resilient infrastructure that protects your business every single day.