Scenario Analysis: an Excavator Cut the Fibre — What Failover Day Looked Like at the Office

This article is a composite, anonymised account of recurring real-world cases; times and measurements are constructed for teaching purposes.

Tuesday morning, on a street in an industrial district, infrastructure digging strays half a metre off plan, and at 10:12 the fibre bundle serving the area snaps. Follow two businesses on that street: a 40-person distribution firm that built a dual-WAN failover arrangement six months earlier, and a similar-sized neighbour running on a single fibre. Same excavation, two different days.

Chronology: the Failover Office That Day

Time	Event	Felt in the office
10:12:04	Fibre cut; the firewall's health checks fail in succession	Nothing yet
10:12:26	Traffic moved automatically to the backup line (a second circuit on different infrastructure)	One refresh on open pages; 15–20 seconds of freeze in a few video meetings
10:13	Automatic alert to IT and to us: "WAN1 down, running on backup"	The dashboard knows — not the users
10:14–10:20	Three active VoIP calls dropped and were redialled; new calls flow over the backup	Six minutes of light turbulence at the PBX
10:25	Fault ticket opened with the provider; regional dig-related outage confirmed	A one-line note sent to management
11:00	Quota protection active: the cloud backup job and large downloads deferred to end of day	Unnoticed
14:37	Fibre repaired; after the stability window, traffic returned automatically (failback)	Nothing — the return was quieter than the switch

End-of-day measurements: switchover 22 seconds, time on backup 4 hours 25 minutes, dropped active calls 3, lost orders 0. Most of the team learned about the outage from the manager's end-of-day summary message.

Meanwhile, Next Door

For the single-line neighbour, 10:12 is when the day ends: no cloud ERP, no email, no card payments. The first half hour goes to modem resets; the provider says "regional fault, being worked on" but gives no ETA. In the afternoon a few staff rescue critical emails over phone hotspots; dispatch paperwork is written by hand, two customers are postponed with "our system is down". The connection returns for everyone at once at 14:37 — but the management that prices those four and a half hours signs the backup-line proposal the following week. The most expensive consultant is an outage you have lived through.

Five Lessons from the Day

• 1. The alert is half the failover: without the switchover notification, the office would not have known it was on the backup, and the primary line's fault might have gone unreported for days. A silent backup lasts exactly until the second outage.
• 2. The diversity principle did its job: had the second line come through the same bundle on the same street, the digger would have cut both. The installation-day question — "do these two circuits share a route?" — paid for itself that morning.
• 3. VoIP expectations were managed in advance: the team already knew "active calls drop and are redialled"; three dropped calls registered as known behaviour, not panic.
• 4. Quota protection was the quiet hero: without the deferral rules, the afternoon cloud backup would have choked the backup line and bred a "the backup is slow too" perception.
• 5. The monthly drill bought confidence: the switchover was not being tried for the first time that day; it ran on the first Friday of every month. The 22 seconds was not a surprise but a repeat of the value in the drill logs.

One Wrinkle, and How It Was Closed

The day was not flawless — honest analysis writes the wrinkle down too: two field devices connecting inbound to head office could not connect for 20 minutes after the switch. Cause: one of them was defined only against the primary line's static IP. The fix went in the same week: critical external access was updated to know both lines' addresses, and the check joined the monthly drill list as a permanent item. That is exactly how failover projects mature: installation + drills + one screw tightened after every real event.

Three Questions for Your Own Office

A manager reading this scenario should ask three questions: If a digger works your street at 10:12 tomorrow, in how many minutes does the office return — can it? Is your backup line (if any) on a genuinely independent route? Who tells you about the switchover — the dashboard, or the users? If the answers are clear, this article was a rehearsal; if not, the first step is a survey.

What Yamanlar Bilişim Provides

Our failover installations are delivered as in this scenario: a diversity-audited second line, automatic switchover plus alerting, quota-protection rules, dual-address definitions for external access, and a monthly drill calendar. For maintenance-agreement customers, line health lives in continuous monitoring; outage day starts and ends on our dashboard, usually before a user notices.

FAQ