Industrial 4G Backup — Automatic Fallback Connection in Less Than 8 Seconds | Eziwan

Why Is an Industrial 4G Backup Essential?

Fiber optics and ADSL are physical infrastructures—they can be disrupted. Roadwork, cables severed by construction equipment, operator equipment failure, or a failure of the local DSLAM: the causes of outages are numerous and often unpredictable. In an industrial setting, a network outage can mean:

• SCADA monitoring outage: loss of visibility into your ongoing processes
• Remote access interruption: your technicians can no longer perform remote maintenance
• Data transmission blocked: meter readings, telemetry, and event logs are not transmitted
• Alerts triggered: monitoring systems generate false network alarms

Industrial 4G backup is the standard solution to this problem. It consists of a 4G LTE cellular connection that remains on standby, ready to take over from the primary connection within seconds.

Dual WAN Architecture: Primary Fiber + 4G Backup

The Eziwan router, in Dual WAN configuration, supports two WAN connections simultaneously:

Role	Link	Typical Throughput	Latency	Usage
Primary WAN	Fiber / ADSL / SDSL	10 Mbps – 10 Gbps	5–15 ms	Normal traffic
Backup WAN	4G LTE	20–150 Mbps	20–50 ms	Backup only

During normal operation, 100% of traffic passes through the primary link. The 4G link is in active standby mode: the router periodically checks its availability but does not use it for production traffic.

If a failure is detected on the primary link, 100% of the traffic is switched to 4G—automatically, without human intervention, in less than 8 seconds.

Fault Detection: Monitoring Mechanisms

The quality of the failover depends on how quickly and reliably a failure is detected. The Eziwan router uses a multi-method detection approach:

Physical link monitoring: loss of Ethernet signal (link down) or loss of DSL synchronization. Instant detection (< 1 second). Covers complete physical outages.

ICMP (ping) probes: Pings sent to reference servers (Google DNS 8.8.8.8, Cloudflare 1.1.1.1) at a configurable interval (5 to 30 seconds). Detects black holes (active physical link but failed routing, which is common during carrier outages).

DNS Probes: DNS queries sent to the operator’s resolvers. Detects DNS service outages that render the Internet unusable even if the physical connection is active.

HTTP Probes: Check the availability of a specific URL (your SCADA server, your cloud). Detects upstream application outages.

Failover is triggered when N consecutive probes fail (configurable: 2 to 5 probes), which prevents failovers caused by transient micro-outages.

Back to the main link: stabilization logic

The logic behind falling back to the primary link is just as critical as the failover itself. Returning too quickly to an unstable link (one that fluctuates between available and unavailable) causes “flapping”—repeated switchover events that disrupt applications more than the outage itself.

The Eziwan router applies a stabilization period before switching back:

• The primary link must be detected as stable for at least 60 seconds (configurable from 30 seconds to 10 minutes)
• The fallback can be configured in automatic mode (fallback as soon as stability is detected) or manual mode (fallback only upon operator action)
• A grace period protects ongoing sessions during the fallback

Maintaining VPN and SCADA Sessions During Failover

In Cold Standby mode (default configuration), a failover results in a change to the WAN IP address. Existing TCP sessions (VPN, SCADA, remote access) are interrupted and must reconnect. For standard industrial applications, the automatic reconnection of VPN and SCADA clients takes 5 to 15 seconds—a process that is generally transparent to operators.

For critical sites that require complete session continuity, two advanced options are available:

VRRP (Virtual Router Redundancy Protocol): Two Eziwan routers (primary and secondary) share a virtual IP address. If the primary router fails, the secondary router takes over the virtual IP address, and traffic continues without interruption. IP sessions are maintained.

Stateful Failover: Synchronization of the connection table between two Eziwan routers. During a failover, the secondary router takes over the exact state of the primary router—including TCP sessions, VPN tunnels, and NAT tables.

Use Cases: Industries Where 4G Backup Is Critical

Continuous-process industries (chemical, petrochemical, food and beverage): SCADA monitoring must remain operational 24 hours a day. A loss of monitoring can lead to a production line shutdown or undetected process deviations.

Energy Distribution (electricity substations, gas and water networks): Energy network operators are subject to regulatory availability requirements. A 4G backup is often required by the concession grantors’ specifications.

Logistics and Warehouses: A WMS (Warehouse Management System) outage halts incoming and outgoing operations. A 4G backup ensures business continuity while the fiber service provider performs maintenance.

Just-in-time production sites: In a just-in-time system, a disruption in EDI exchanges with suppliers can halt both procurement and production.