The cost of network connectivity resilience that nobody budgets for
When business leaders discuss network investment, the conversation usually centres on bandwidth, speed and monthly line costs. It rarely accounts for what happens when the connection drops. Research consistently puts the cost of unplanned downtime at thousands per minute once you factor in lost transactions, idle staff, delayed logistics and regulatory exposure. Nearly half of IT and business leaders believe a network outage would have a major negative impact on their operations, yet a significant proportion of those same organisations have never verified whether their backup connectivity would actually function during a real outage.
That gap between perceived protection and actual resilience is where the real risk lives. We saw it play out globally in July 2024, when a single CrowdStrike software update crashed an estimated 8.5 million Windows systems simultaneously. Airlines grounded flights. Hospitals reverted to paper records. Banks couldn’t process payments. The cause wasn’t a carrier fault or hardware failure. It was a third-party software dependency that nobody had identified as a single point of failure.
The lesson wasn’t that technology is unreliable. It was that resilience cannot be assumed. It has to be designed.
The redundancy illusion
This is something we encounter regularly. A business has two internet connections from different providers and assumes that constitutes failover. It often doesn’t.
Across both the UK and South Africa, we routinely find environments where dual connections are delivered over shared last-mile infrastructure. A single physical fault, a damaged duct or a street cabinet failure, affects both services simultaneously. The business has paid for redundancy but purchased an illusion.
In other cases, a secondary line exists but has never been configured for automatic failover. When the primary drops, the switchover needs manual intervention at precisely the moment when the internal team is most stretched. And sometimes the backup was configured years ago and never tested under real conditions. When a genuine outage hits, they discover the backup can’t handle the bandwidth their critical applications actually need.
Our approach is straightforward: redundancy that hasn’t been tested isn’t redundancy. It’s an assumption, and assumptions fail at the worst possible moment.
Where modern architecture helps with business continuity planning — and where it doesn’t
The shift to cloud-hosted applications has fundamentally changed the stakes. When your ERP, your payment platform or your logistics system lives in the cloud, a network outage doesn’t just inconvenience users. It makes the application entirely inaccessible. There’s no local fallback.
This is where SD-WAN has become genuinely valuable. Rather than treating all traffic equally, it lets businesses prioritise mission-critical systems during periods of network strain or failover. Your ERP and VoIP get preferred treatment whilst less time-sensitive traffic is handled appropriately.
But SD-WAN is an intelligent routing layer, not a resilience solution in isolation. Built on top of poorly diversified circuits, without proactive monitoring or tested failover logic, it won’t prevent an outage from affecting the business. It will simply manage traffic more elegantly until it can’t.
The most resilient architectures we design combine SD-WAN’s routing intelligence with genuine carrier diversity, physical infrastructure separation and continuous performance oversight. Each layer addresses a different failure mode. Together, they create something no single technology delivers alone: operational continuity.
The governance gap
During a connectivity disruption, one of the biggest contributors to extended downtime isn’t the fault itself. It’s fragmented accountability. When multiple ISPs, hardware vendors and cloud providers are involved, internal teams find themselves coordinating escalations across several organisations simultaneously whilst trying to maintain business operations.
This led to a principle we’ve applied consistently over the years: a resilient architecture must include a single experienced team that owns the entire connectivity ecosystem. Carrier relationships, routing policies, performance optimisation, failover testing, incident escalation and ongoing reporting. Our Trusted Response Centre exists precisely because without that ownership layer, even technically sound redundancy gets compromised by slow resolution and unclear responsibility during a crisis.
The organisations that recover fastest from disruption are rarely those with the most sophisticated infrastructure. They’re the ones that know exactly who is responsible and have already agreed how that responsibility works under pressure.
Five questions worth asking before your next board meeting
If any of these generate uncertainty, treat that uncertainty as a signal:
- Are your primary and secondary connections physically separated at infrastructure level, or do they share last-mile infrastructure that a single fault could affect simultaneously?
- Is your failover fully automatic, and when was it last tested under real conditions?
- Are mission-critical applications prioritised during traffic strain and failover, or does all traffic compete equally?
- Would performance degradation be detected before your staff or customers experience it?
- Is there a single accountable team managing your entire connectivity environment, or is responsibility distributed across multiple vendors with no single point of escalation?
The organisations that recover fastest from disruption are rarely those with the most sophisticated infrastructure. They’re the ones that know exactly who is responsible — and have already agreed how that responsibility works under pressure.
Designing stability into the background
When resilience is designed thoughtfully, monitored continuously and governed with clarity, disruption becomes largely invisible to the business. Systems continue operating. Teams remain productive. Customers experience continuity. And leadership retains the strategic confidence to grow.
That’s not an IT outcome. That’s a business outcome.
