4G/5G failover, the testing nobody does.
How business internet failover actually works. What survives an outage. Why most installed failover doesn't trigger when the broadband goes down. The 10-minute quarterly test that prevents the day-of-outage disaster.
- 🔄 How failover actually works
- 📞 What survives the cutover
- ⏱ Detection and switch timing
- 🚨 The silent failure modes
- 🧪 Quarterly testing procedure
- 📡 Multi-network SIM technology
- 🛡 Antenna and signal optimisation
- 💷 Cost vs leased line resilience
What 4G/5G failover actually does during an outage.
4G/5G business internet failover uses a router with built-in mobile modem (or a separate cellular gateway) that holds a SIM card for one or more UK mobile networks. The router monitors the primary internet connection (usually FTTP or leased line) continuously. When the primary fails — cable cut, carrier outage, equipment failure, power problem upstream — the router detects the failure and routes all internet traffic through the mobile path instead.
What "detection" actually means: the router doesn't just wait until the WAN port shows "no link" — that's too slow. Proper business failover routers use active monitoring: pinging known reliable endpoints continuously, watching for consistent packet loss, and checking specific service paths (e.g. is the VoIP provider's SBC reachable?). When the monitoring confirms the primary is genuinely down (not just a transient glitch), it triggers failover.
What "failover" actually means: the router updates its routing tables to send all outbound traffic through the mobile interface. Existing TCP connections (file uploads, video streams) generally break and need to re-establish. Stateless connections (VoIP calls, browsing) reconnect quickly through the new path. Applications that don't notice the path change continue running. Applications that hold long-lived TCP connections may need a manual reconnect.
When primary returns: the router detects the primary is back, waits a brief stabilisation period (often 60-120 seconds to confirm it's genuinely stable, not flapping), and switches traffic back. Same mechanics in reverse.
The honest breakdown of what works during failover.
VoIP calls in progress
Survives: with fast failover (5-15 seconds), calls in progress survive a brief audio glitch and continue. Without fast failover, calls drop and need to be redialled. The cutover is the most critical moment — proper QoS configuration and active monitoring are required for clean preservation.
Card payment terminals
Survives: card terminals reconnect via the failover path within seconds. For retail and hospitality, this is often the most important survival case — every minute of payment outage = lost trade. Standard practice: terminals tested specifically during failover testing.
Email and cloud documents
Survives: staff can keep reading email, working on cloud documents (Google Docs, Microsoft 365), and accessing cloud applications. Performance may be slower over 4G than over fibre but functionality is preserved.
Video conferencing
Partially survives: 5G failover handles video calls reasonably well. 4G failover may have quality issues, particularly with multi-participant calls. Single 1:1 video calls usually work; large all-hands meetings may degrade.
Cloud backups in progress
Doesn't survive cleanly: backup operations holding long-lived connections break during failover. Some backup software resumes automatically; some needs manual restart. Scheduled backups during peak business hours are best avoided regardless.
Large file uploads
Doesn't survive: large file transfers (multi-GB uploads to cloud storage, video file transfers) break at cutover. Failover bandwidth is limited; mass uploads aren't realistic over mobile. Plan critical uploads outside failover periods.
Why most "installed" failover doesn't actually work.
The dirty secret of business failover: a large percentage of installed failover deployments don't actually trigger correctly when the primary goes down. The failover hardware is in place, the SIM is active, the router shows "ready" — but on the day of the outage, the failover doesn't work as expected. Common reasons:
SIM not provisioned correctly. The SIM works for the initial activation test but the data plan doesn't include the kind of usage actually needed during a real outage. Some SIMs work for hours-of-light-use but get throttled or suspended when an actual office routes all its traffic through them.
Mobile signal too weak. The installer tested with a smartphone showing 4 bars of signal at the install location — but the router's external antenna is shielded by the metal of the rack or the building structure. When the router actually tries to push office-volume traffic through the mobile connection, signal quality is insufficient.
Router monitoring settings wrong. Default monitoring may only trigger failover when the WAN port shows "no link" — but most internet outages are partial (high latency, intermittent connectivity, DNS failures) rather than complete cable cuts. The router doesn't see the partial failure, doesn't trigger failover, and the office sits unable to work while the failover sits unused.
Routing tables not updated. The router knows failover is active but specific traffic types are still configured to use the primary path. VoIP traffic might fail over correctly while web traffic doesn't, or vice versa. Inconsistent state means "some things work, some don't" during outages.
Failover triggered but bandwidth saturated. A 100-person office trying to push everything through a 4G connection saturates immediately. No prioritisation between essential traffic (VoIP, card payments, email) and non-essential (streaming, downloads). Result: everything is slow, key services don't get bandwidth. Proper QoS on the failover path prevents this.
All these failure modes are invisible until tested. Hence:
How to actually test your failover works.
The test that prevents the day-of-outage disaster. Should be done quarterly minimum. Costs 10 minutes. Catches almost every silent failure before it matters.
Step 1 — Schedule the test. Outside peak hours but within business hours so staff can observe the impact. Inform staff that a brief blip will occur. Have someone ready to test specific services.
Step 2 — Disconnect the primary. Physically pull the WAN cable at the router, or disable the WAN port in router configuration. The clock starts now.
Step 3 — Observe failover trigger. Time how long until the router shows traffic flowing through the secondary path. Target: under 15 seconds. Acceptable: under 60 seconds. Concerning: anything slower.
Step 4 — Test critical services.
- Make a VoIP call and confirm it works clearly
- Process a £0.01 test card transaction
- Open a cloud document, edit it, confirm changes save
- Send and receive an email
- Browse to a website that's not heavily cached
Step 5 — Measure bandwidth. Run a speed test through the failover path. Confirm bandwidth is reasonable for the office size.
Step 6 — Restore the primary. Reconnect the WAN cable. Time how long until the router switches back. Verify the switch-back doesn't cause issues with in-progress operations.
Step 7 — Document results. Note failover trigger time, what worked, what didn't, any specific issues. Compare against last quarter's results — degradation between tests indicates problems developing.
Most businesses never do this. Telexico-managed failover includes quarterly tests with documented results as standard.
The unsung hero of reliable mobile failover.
Single-network SIMs (e.g. an EE-only SIM) work fine when EE has good coverage at your install location. They fail when EE's coverage degrades, or when EE itself has an outage in your area.
Multi-network SIMs (sometimes called "steered" or "operator-agnostic" SIMs) connect to any of the UK mobile networks (EE, Vodafone, O2, Three) and automatically switch between them based on which network has the strongest signal at the location, or which network is functioning properly during an outage.
This matters for failover specifically because: (a) the mobile network you tested with at install time may not be the best network at the exact moment of an outage, (b) the UK mobile networks occasionally have regional outages — having access to multiple networks dramatically improves the chances that failover actually works when needed, (c) multi-network SIMs auto-optimise signal quality over time as nearby cell tower coverage changes.
Multi-network SIM cost is similar to single-network SIMs at typical UK business volumes. Almost no reason not to use them for business failover.
The cheapest performance upgrade most installs miss.
Most 4G/5G routers have built-in or short-stub external antennas. These work in environments with strong mobile signal but underperform in marginal-signal locations — particularly inside large buildings, behind metal cladding, or in basements where most comms rooms live.
External directional antennas mounted on the building exterior, pointed at the nearest cell tower, dramatically improve signal quality. The performance difference between an indoor stub antenna and a roof-mounted directional antenna can be 20-30dB — the difference between barely-working failover and full-speed failover.
Installation: typically a small outdoor antenna ~£100-300, mounted on the roof or external wall, with low-loss coax cable run to the failover router. Total installation cost £200-600 including labour. The performance improvement justifies the cost for any failover deployment in a marginal-signal location.
Test your signal quality before installation. If a smartphone at the install location shows 1-2 bars of signal, external antenna is essentially mandatory. If smartphone shows 4-5 bars throughout the building, internal antennas may be sufficient.
Failover delivers 80%+ of leased-line resilience at 10% of the cost.
The honest economic comparison: FTTP business broadband + 4G/5G failover vs leased line for typical UK SME resilience needs.
FTTP + failover route:
- FTTP business broadband: ~£60/month
- 4G/5G failover with multi-network SIM: ~£25-35/month
- Total monthly: ~£85-95
- 3-year cost: ~£3,100-3,400
Leased line route:
- 100Mbps leased line urban: ~£300/month
- Plus install: ~£1,500 one-off
- 3-year cost: ~£12,300
Resilience comparison: the FTTP + failover route handles 95%+ of business operational scenarios — VoIP calls preserved, card payments live, cloud apps working. Where it falls short: heavy upload operations during outages (not realistic on mobile bandwidth), formal SLA-backed uptime guarantees (the leased line has them; failover relies on dual-provider availability).
For most UK SMEs the FTTP + failover route is the economically rational choice. For operations where 30 minutes of additional uptime per year is worth £3,000 per year, the leased line maths works. For operations where it isn't, failover wins.
Review my current setup.
Not ready to switch yet? Send us your current contracts, bills, or photos of your existing equipment. We'll review what you have, what you're paying, and where you could simplify, consolidate or improve — without any pressure to buy anything from us.
We benchmark your existing broadband, phones, mobile and IT against current UK market pricing and what your business actually needs.
Real engineer review of your current connectivity, voice setup, WiFi, security and continuity — strengths, gaps, and where you're overpaying or underprotected.
We tell you honestly what's available at your postcode — FTTP, leased line, alt-net carriers — and which makes commercial sense for your operation.
For businesses still on ISDN or aging on-premises PBX — an honest cost-and-feature comparison before the 2027 BT switch-off forces a rushed decision.
No hard sell. No fixed package pressure. If we're not a better fit, we'll tell you straight — and recommend what is.
Frequently asked questions
What is 4G/5G failover and how does it work?
4G/5G failover provides a secondary internet connection via the UK mobile networks (EE, Vodafone, O2, Three) that activates automatically when your primary broadband or leased line goes down. A failover-capable router monitors the primary connection continuously and switches to the mobile data connection within seconds of detecting a failure. When the primary comes back online, the router switches back. The mobile connection typically delivers 50-300Mbps in good coverage areas — enough to keep VoIP, card payments, and cloud apps working through an outage.
What survives an internet outage with proper failover?
With well-engineered failover: VoIP calls in progress get preserved (the router fails over within seconds; calls briefly stutter then continue), card payment terminals continue working (they reconnect via the mobile path), cloud applications continue working (slower but functional), staff can keep working on cloud documents and emails. What doesn't survive: heavy bandwidth operations like cloud backups, video calls may have quality issues, large file transfers. The mobile connection has much lower bandwidth than the primary, so the experience is 'survival, not full performance'.
How fast does failover actually trigger?
Depends on the configuration. Aggressive failover (1-5 second detection): switches over before users notice anything beyond a brief audio glitch. Standard failover (30-60 second detection): noticeable interruption but recoverable. Slow failover (default consumer router setting, often 2-5 minutes): basically useless for keeping calls and payments live. Proper business failover should detect and switch within 5-15 seconds — measured by how long the primary connection is dead before traffic is flowing through the secondary path.
Why doesn't my failover actually work when broadband goes down?
The most common reason: it was never tested. Failover gets installed, the router shows the secondary connection is active, but nobody ever pulls the primary cable to verify the failover actually triggers correctly. Common silent failures: SIM card not provisioned correctly (works initially, dies after first usage spike), mobile signal too weak at the install location (worked when tested with a phone but the router's external antenna is shielded), router monitoring settings too lax (doesn't trigger failover on partial degradation), routing tables not updated so traffic doesn't actually use the failover path. Quarterly testing catches these before the day they matter.
How do I test my failover?
Disconnect the primary internet at the router — physically pull the cable or disable the WAN port in router configuration. Observe what happens over the next 30 seconds. Make a test phone call and confirm it continues working. Process a test card transaction. Open a cloud document and confirm changes save. Refresh a cloud app and confirm it loads (slower is acceptable; broken is not). Restore the primary connection and verify the router switches back. This whole test takes 10 minutes and tells you whether your failover actually works. Most businesses never do it. Telexico-managed failover includes quarterly testing as standard.
Is 5G better than 4G for business failover?
Yes where available — 5G delivers higher peak speeds (100-500Mbps in good coverage vs 50-150Mbps for 4G), lower latency, and better performance in congested cells. For failover use, the practical difference is whether the failover path can carry video calls (yes on 5G; sometimes marginal on 4G) and how many simultaneous users it supports. In areas without 5G coverage, 4G is still sufficient for keeping VoIP, card payments and core cloud apps running through an outage. Most Telexico failover deployments use multi-network SIM technology that automatically uses 5G where available and falls back to 4G where not.
What's the cost of business 4G/5G failover?
Typical UK 4G/5G failover deployment: hardware (router with built-in modem or separate cellular gateway): £300-800 one-off, or rolled into monthly fee. Monthly service: typically £20-40/month including SIM data allowance (usually 500GB-unlimited). Compared to a leased line at £250-500/month for similar resilience, failover at £25/month is dramatically cheaper. For most UK SMEs, FTTP + 4G/5G failover delivers nearly the same operational resilience as a leased line at 10-15% of the cost.
Tailored around your business.
Send us your current setup. We'll review what you have, what you pay, and where we can simplify, consolidate or improve it — no hard sell, no fixed-package pressure.