The recent nationwide Optus outage in Australia sent the media into an absolute meltdown. Trains stopped. Emergency services choked. The collective commentary reached a predictable, lazy consensus: "This is a systemic failure of corporate responsibility, and we need government-mandated, bulletproof redundancy so this never happens again."
That narrative is not just wrong. It is economically illiterate.
The public reacts to a major network collapse the same way they react to a plane crash—with irrational horror fueled by a complete misunderstanding of risk engineering. I have spent two decades building and auditing core network architectures. Here is the uncomfortable truth nobody in upper management wants to say out loud: building a telecom network that never goes down is financially impossible, and trying to do so will bankrupt the very consumers crying for it.
We do not have a technology problem. We have an entitlement problem.
The Myth of the Unbreakable Network
When a network routing error drops 10 million people into the digital dark ages, the immediate demand is for "total redundancy."
This stems from a fundamental misunderstanding of high-availability engineering. In telecommunications, availability is measured in "nines." Three nines ($99.9%$) means roughly nine hours of downtime a year. Five nines ($99.999%$) means about five minutes of downtime a year.
To move from three nines to five nines does not cost double; it costs exponentially more. It requires entirely separate physical fiber paths, duplicate hardware from competing vendors, and separate power grids. If you force carriers to guarantee absolute uptime, your monthly phone bill will triple.
Worse, complexity is the ultimate enemy of security. When you stack redundancy upon redundancy to appease regulators, you create a fragile, over-engineered monstrosity.
The Redundancy Paradox: The more backup systems you introduce, the more configuration points you create. The more configuration points you create, the higher the probability that a routine software update will cause a cascading, unpredictable systemic failure.
Imagine a scenario where an engineer pushes a routine Border Gateway Protocol (BGP) update. In a simple network, it fails locally, and you roll it back. In an aggressively redundant network with automated failovers, that corrupted update propagates across multiple layers simultaneously, tricking the backup systems into replicating the error. The safety nets themselves become the executioner. This is exactly how the largest digital blackouts in history happen.
Dismantling the Panic
Let us address the questions that fill search engines every time a tower goes dark. The premise of these questions is almost always flawed.
Why do emergency calls fail during a network outage?
The public assumes that if Carrier A goes down, their phone should just seamlessly jump to Carrier B to dial triple zero or 911.
The brutal reality is that emergency roaming requires structural interoperability at the deepest layers of national infrastructure. If a core network routing table is completely wiped or corrupted, the device cannot authenticate itself anywhere. It becomes a brick.
We can spend billions trying to build a secondary, state-owned emergency wireless layer, but that capital would be stripped away from healthcare, roads, and education. We accept a baseline level of risk in every other facet of life. We do not build ten-lane highways just in case there is a fender bender on the main road. Why do we expect it from data packets?
Why can't trains run without a commercial telecom provider?
When transport systems grind to a halt because a commercial carrier drops connection, the anger is directed at the carrier. It should be directed at the transit authorities.
Relying on a consumer-grade, commercial telecom network to run critical public transport infrastructure is an act of sheer corporate negligence. Mass transit systems should operate on dedicated, air-gapped, private radio networks (like GSM-R or dedicated LTE slices). If a train system paralyzes a city because a commercial provider had a bad day, the transit system failed its own basic risk assessment. The carrier is just a scapegoat for cheap public procurement.
The Real Cost of Absolute Reliability
Let us look at the heavy hitters. Companies like AWS, Google, and Microsoft spend billions annually on infrastructure. Yet, even they experience spectacular, multi-hour outages.
If the richest entities on Earth cannot buy 100 percent uptime, a regional telecom operator certainly cannot.
If we bow to the political pressure of enforcing strict financial penalties for every second of downtime, carriers will simply stop innovating. They will freeze their architectures. They will refuse to implement necessary security patches or upgrade to faster protocols because the risk of a configuration error becomes a corporate death sentence. You will be stuck with ancient, slow, but "stable" technology forever.
There is a downside to this contrarian view, of course. Accepting that networks will fail means accepting that businesses will lose money, some transactions will fail, and lives will occasionally be disrupted. It is cold. It is clinical. But it is honest.
Stop Demanding Fixes, Start Designing for Failure
The solution to the next inevitable blackout is not to drag telecom executives before a senate committee and demand they promise it will never happen again. They will lie to you, say what you want to hear, and pass the compliance costs back to your wallet.
Instead, the responsibility must shift to the enterprise and the individual.
- Dual-SIM is cheaper than legislation: If your business relies on connectivity to survive, buy a secondary SIM card from a rival network. It costs pocket change compared to the losses of a six-hour outage.
- Air-gap critical systems: Governments must legally mandate that critical utilities—water, power, emergency signaling—cannot share core routing architecture with commercial consumer traffic.
- Embrace the analog fallback: Merchants need to stop throwing their hands up when EFTPOS goes down. Keep a manual processing system or a cash reserve on hand.
Stop treating the internet like oxygen. It is a highly volatile, human-made utility held together by code, fiber optic cables buried in the dirt, and underwater lines chewed by sharks. It is going to break. Deal with it.
