In the spirit of ruffling some feathers, let’s get straight to it—my introduction to SD-WAN wasn’t born out of curiosity or a fascination with new technology. It was a necessity, driven by the sheer inefficiency of traditional network management. My journey started while I was on a contract at an ISP up the road, tasked with analysing costs and introducing optimisation strategies. What I uncovered was a perfect storm of human-driven processes, manual interventions, and an overwhelming reliance on spreadsheets.
The ISP’s Operational Nightmare | A Sea of Tickets and Redundant Links
The ISP in question had a fundamental problem—its services were essentially managed by downloading and uploading Excel spreadsheets. Everything required human oversight, and this reliance on manual processes inflated operational costs. One of the key metrics I examined was headcount per line, and let’s just say their numbers were far worse than any competitor. The efficiency ratio was abysmal.
But the real eye-opener came when I dug into their support ticket queues and fault categorisation. The vast majority of support tickets were link failures. The sales team seized this data and used it to convince customers to buy redundant links. On paper, it seemed like a solid strategy—after all, redundancy should reduce downtime, right?
Wrong.
Instead of stabilising the network, redundant links increased the number of support tickets. The sheer volume of issues overwhelmed the support staff. There simply weren’t enough hours in the day for engineers to sift through the growing mountain of tickets. And then, someone had the bright idea to introduce a script to auto-close tickets in bulk.
You can guess what happened next. Clients simply reopened their tickets. The ISP had built a self-perpetuating loop of inefficiency.
The Human-Intensive Failover Problem
A deeper investigation revealed the root cause—the failover solution itself was manual. The ISP had deployed dynamic routing protocols to handle failover, but engineers quickly realised that failovers caused unpredictable flapping. In many cases, automated failovers created more instability than they solved. The easiest way to ensure site stability? Disable automated failover entirely.
This meant that when a link actually failed, a network engineer had to manually log into the router via the backup path and redirect traffic. This was all happening while support teams were drowning under a staggering 8,000 tickets per day—in a network with around 45,000 customer links.
Enter SD-WAN | A Lightbulb Moment
I started researching SD-WAN solutions and even examined NTT’s short-lived variant. The premise was clear: instead of relying on traditional routing protocols to determine link failure, SD-WAN uses an overlay network that actively monitors performance metrics in real-time. When a link degrades or fails, traffic is instantaneously rerouted—no need for human intervention.
It was a revelation.
SD-WAN wasn’t just another networking tool. It was, at its core, an automation engine, removing the human bottlenecks that plagued traditional ISP networks. The ability to intelligently orchestrate failovers, monitor performance, and self-heal in real-time was precisely what the ISP needed.
The Road Not Taken… and Still Not Taken
Unfortunately, the ISP was soon folded into its parent company, and my vision for a fully automated SD-WAN-driven network never materialised. But the lesson was clear: traditional failover methodologies are fundamentally broken. Many ISPs and enterprises are still making the same mistake today—relying on manual interventions and reactive troubleshooting instead of embracing intelligent automation.
To this day, I see the same debates and discussions in the industry, proving that for many, the problem remains unresolved. But for those who have embraced SD-WAN, particularly advanced solutions like Fusion’s SD-WAN, the benefits are clear:
Automated failover that works—no manual intervention required.
End-to-end visibility with real-time analytics—no more flying blind.
Self-healing networks that reduce support tickets—because customers don’t need to report failures if they never experience them.
For me, SD-WAN was never just about connectivity. It was about automation, efficiency, and eliminating unnecessary complexity. The technology has matured, but its core principle remains the same: stop managing networks like it’s 1999, and let automation take the wheel.
And if your ISP is still managing failover with a script that bulk-closes tickets? Well, it’s time to move on.
