On 13 October 2025, Vodafone experienced an outage that left many mobile and broadband customers with full signal — but no Internet.
External BGP monitors showed that the carrier's main Autonomous System (AS) briefly disappeared from the global routing table. In other words, the Internet stopped seeing routes to parts of Vodafone's network.
So, while devices stayed connected to nearby cell towers, their data had nowhere to go beyond them. The radio network was fine; the path through the carrier's routing system wasn’t. No way out, no way back.
When we checked the live data sessions on our IoT estates during the same window, our Multi-network SIMs connected to Vodafone kept passing data.
Here’s why:
It’s a small architectural distinction, with big operational consequences.
And it’s a lesson in what a built-for-purpose IoT Network really means for the channel.
Not to mention, what the next decade of connectivity will look like: faster, broader, and far more interdependent.
Incidents like this are becoming more visible because the way networks are built — and relied on — has changed dramatically.
Twenty years ago, a carrier’s network was a self-contained ecosystem. It owned the towers, the core, the routing, and the customer relationship. Outages were localised, and their impact stopped at the operator’s borders.
Today, networks are layered, virtualised, and globally intertwined. A misconfiguration in one place can cascade through multiple providers in seconds.
Four key forces explain why resilience now depends on architecture, not size:
The Internet’s routing fabric (BGP) operates on trust. When key routes are withdrawn, entire networks can drop out of global reachability.
Retail, wholesale, and IoT traffic now often share virtualised systems. This is efficient, but means that a single routing policy change can affect multiple customer types at once.
Control functions that used to sit in physical network nodes now often live in the cloud. When a cloud region fails, it can take part of a network’s routing logic with it.
From video calls to vehicle telematics, every byte can travel through the same infrastructure. Managing this diversity introduces latency, congestion, and new failure points.
For end users, all these layers can collapse into one experience: “it’s down.”
For resellers, it means a carrier’s reliability no longer equals your reliability.
When it comes to IoT and business critical applications, resilience has to be designed beyond the carrier layer — in the routing and breakout decisions that determine where traffic exits the network.
When people talk about resilience, they usually mean coverage: more networks, more redundancy, more signal strength.
Coverage is vital — it’s what keeps devices connected to something. But coverage alone doesn’t guarantee continuity.
Modern resilience is layered:
Coverage (radio access):
Keeps devices connected to a network but can’t protect against routing or core failures.
Capacity (performance):
Keeps data moving efficiently, handling congestion and throughput, but doesn’t prevent disconnection when a carrier’s routing fails.
Routing:
Decides how and where traffic exits the network, and determines continuity during outages or policy faults, but requires independent management to be effective.
In other words:
Coverage gets you on the network.
Capacity keeps your data moving.
But routing is what can keep you online when parts of a carrier’s network fail.
So while multi-network coverage still forms the foundation of resilience, independent routing adds another layer — one that keeps customers online even when a carrier’s routing or core layer fails.
Together, they form multi-path resilience: coverage for connection, routing for continuity.
And in an era of shared cores, virtualised routing, and multi-carrier systems, that routing layer has become the real test of resilience.
It’s important to note that carriers' routing is world-class.
But it’s designed for mass-market consumer traffic.
It’s built for reach and efficiency, prioritising throughput, consistency, and cost-efficiency.
What makes independent routing better for IoT and channel partners specifically is that it’s designed for control and resilience.
It prioritises uptime, flexibility, visibility, and data control for managed IoT and enterprise services.
So, while carriers give you global reach, independent routing gives you the ability to decide how that reach is used, where traffic exits, and what happens when something breaks.
As with all modern connectivity choices, it’s about designing solution-fit.
For telecom/IT resellers, MSPs, and IoT solution providers, the Vodafone outage is a reminder that how your provider handles routing, redundancy, and breakout directly affects the service your customers experience.
Below, we’ve broken down the implications — and how to turn them into sales and marketing advantages.
In most wholesale models, data leaves the network through the carrier’s Internet gateway. When that gateway fails, every service downstream fails too.
But providers who manage their own APNs and breakout can bypass this dependency. Data may enter through a carrier’s radio access network, but it exits through an independently controlled route.
That’s why our IoT SIMs kept passing data across the Vodafone outage window, even while many other services dropped.
In the traditional mindset, reliability meant sticking with a household-name carrier. But scale no longer guarantees resilience.
Large carriers are primarily optimised for throughput and efficiency, not for isolating every tenant or traffic type. Enterprise and consumer traffic often share routing layers, so a single configuration issue can affect both.
By contrast, a cellular ISP that operates across multiple carriers but runs its own core, routing policies, and Internet breakout can isolate faults and reroute around them.
Independence doesn’t replace the carrier, but it does add a protective layer above them.
Better yet, this story gives you a way to reframe “multi-network” as something deeper and more defensible.
Right now, most resellers say:
“We sell multi-network SIMs, so your devices always connect.”
Now you can evolve it to:
“We work with a provider whose network stays online even when major carriers go down. So you get more than multi-network coverage — you get multi-path resilience.”
That’s a powerful, business-level message that doesn’t require explaining APNs or routing.
Customers rarely care about the underlying tech — until it fails. Then, it defines their perception of value.
That’s why the outages offer such clear talking points. They shows the difference between connectivity bought as a commodity and connectivity designed as an operational safeguard.
Don’t shy away from numbers: a few minutes of downtime across dozens of connected sites can equate to thousands in lost productivity. That’s where resilience becomes ROI.
Customers trust real-world evidence over technical promises.
When outages happen — and they will — they create case studies for partners who stayed online. That’s a rare and credible differentiator in a market where most sales decks sound identical.
That simple line turns an event into a story about foresight.
Connectivity architecture might seem like a technical detail, but its commercial impact is direct and measurable.
Each time a major outage happens, the partners who rely on single-carrier infrastructure spend hours firefighting, issuing credits, and reassuring customers.
Partners who operate with a provider who independently manages their network keep selling while others are troubleshooting.
Every outage you don’t suffer becomes a reputational asset.
The carrier incident or outage won’t be the last test of network resilience.
The next wave of change is already reshaping how routing, orchestration, and control interact:
Each shift deepens the same question: who owns the path out?
The differentiator won’t just be multi-network coverage — it’ll be multi-path resilience too, where traffic has multiple exits, multiple failovers, and one consistent standard of uptime.
If you’re a reseller or MSP, here's your checklist to get started:
_white.svg)
