Backup internet
for a business

It depends on one very specific question, not on the size of the premises: if, the moment the line goes down, you can no longer take payments, answer customers or open the door, then yes, you need it. A countertop POS with its own line, a card terminal that runs over the internet or a cloud POS all stop working as soon as the fibre drops, without warning and with no margin. The same goes for email if it's your only channel for orders, or for a cloud booking system. The useful criterion isn't how many square metres the business has, it's how many hours you can last without a connection before that turns into lost money or a customer going elsewhere. If the answer is "none", a backup isn't a technical luxury, it's an insurance policy that costs less than the first real outage you'd otherwise have to sit through without it.

Only if it genuinely enters the building by a different physical route, and that needs checking rather than assuming because the logo is different. Many operators lease sections of network from one another, and two contracts under different names can end up coming in through the same connection, the same pole or the same street cabinet. The day a workman digs up that trench or that cabinet fails, both lines go down at once, even though the bills come from two different companies. To find out whether your second contract is any use, ask where it physically enters the building: if it shares an entry point with the main line, it isn't a second path, it's the same exposure under another name. What really diversifies is a different medium — a radio link, mobile or satellite — because none of them depends on roadworks leaving your cable alone.

As a one-off stopgap, yes: tethering from your phone will do for checking email or making a call while the outage is fixed. As a business backup plan it doesn't hold up, for two reasons that are about process rather than coverage. The first is that it depends on someone being physically there, remembering, and knowing how to switch it on; if it happens out of hours or that person isn't around, the business stays down regardless. The second is that the POS, the cameras or the office email don't connect to anyone's phone by themselves: you have to go device by device changing the network, something nobody does under pressure with customers waiting. The backup that works is one that doesn't rely on a human reacting in time: a device that detects the outage and switches path on its own, without anyone having to remember anything.

It's the device that turns having "a second line" into having a genuine backup, because without it the second line is a cable nobody plugs in until the problem has already happened. It continuously monitors the main line, and as soon as it detects that it has stopped responding, it automatically moves all the premises' traffic onto the second connection, without anyone having to touch anything or be there. When the main line comes back, the router detects that too and moves the traffic back, again on its own. The office devices don't notice the change of path because they keep talking to the same router and the same internal network; the only thing that changes is the route that traffic takes to the outside world. It's the piece that stops a fibre outage turning into a distress call to whoever happens to be in the shop that day.

Yes, and it fits particularly well where the problem isn't just the fibre but that the area's fibre and mobile coverage share the same underlying infrastructure, which happens more often than you'd think on industrial estates or with shared cabinets. Because it arrives by satellite, it doesn't depend on any terrestrial connection: not the trench in the street, not the cabinet on the corner, not the local mobile mast, so an outage that takes down fibre and mobile at the same time doesn't affect it. What it asks in return is a clear view of the sky for the dish and somewhere to power it, two conditions that need checking before counting on it as plan B, not afterwards. Where those conditions are met, it's one of the most independent backups there is, precisely because it shares no physical section with the network it has to stand in for when the outage hits.

No, and that's exactly the point that most reassures people who set one up thinking they need to duplicate their expensive contract. The backup only has to keep the essentials going while the outage is fixed: the POS taking payments, email coming in and going out, a video call not cutting off mid-sentence. None of that requires the advertised speed of the main line, which in any case is rarely used to the full day to day. What is non-negotiable has nothing to do with megabits: switching from one line to the other must be automatic, with no one having to step in, and the second line's path must not depend on the same infrastructure that fails on the main one. A slow but reliable, automatic backup is infinitely better than a fast one that needs someone to switch it on by hand at the worst possible moment.

Yes, and that's exactly what makes the switchover unnoticeable to everyone except the router. The dual-WAN device sits between the internet connection and the rest of the installation, so the Wi-Fi, the internal cabling, the POS and the cameras keep talking to exactly the same network as always; the only thing that changes is which exit that traffic takes to the outside. Nobody has to reconnect to another network or remember a different password on the day of the outage. This matters more than it seems, because the usual alternative — a second "emergency" Wi-Fi network — almost never works on the day it's needed: nobody has it written down, nobody remembers the password, and devices like the POS don't connect by themselves to a network they don't know. Keeping a single internal network with two exits is what makes the backup ready without anyone having to prepare anything that day.

Properly configured, the switchover is practically instantaneous: the router detects that the main line has stopped responding and diverts the traffic within seconds, not minutes. In practice it shows up as a brief blip — a page that takes a moment longer to load, a video call that freezes for an instant and recovers by itself — rather than a loss of service. The POS usually doesn't even show an error if the switchover is quick, because the time without a connection is shorter than it takes to notice. What should be ruled out from the start is any solution that depends on someone noticing the outage, going to the comms cupboard and swapping a cable by hand: that's minutes, not seconds, and in that time a sale has already been lost or a call dropped. If your plan B needs human intervention to kick in, it isn't switching over, it's waiting for someone to remember it.