What a backup path really means on a live stream

The term, in one sentence

On a live stream, a backup path is a second, separate route that carries the same programme to the audience if the first route fails. It is not a recording and it is not a spare laptop in a bag. It is a complete chain, running at the same time as the main one, from the point where the picture leaves the gallery to the platform that delivers it to viewers.

I hear the phrase used loosely on shows, so it helps to be precise. A path has three parts: an encoder that turns the programme into a stream, a network connection that carries it out of the venue, and an ingest point at the platform that receives it. A true backup path has its own version of each, so one failure cannot take out both.

When the three parts are separate, the audience sees a short glitch at worst. When they share a part, you have a single path with extra cables, and the crew finds that out at the worst moment.

Why one route is never enough

Most live stream failures are dull. A venue router restarts, a cable is knocked loose behind a rack, an upload line slows because someone else on site is sending large files, or a platform ingest server drops the connection for a moment. None of these is dramatic, and none can be predicted from a rehearsal that went well.

That is the reason a backup path exists. We do not plan for the stream that works on the day; we plan for the five minutes in which one thing does not. A single route means that one ordinary fault becomes a dark screen, and viewers rarely wait for it to come back.

Events and festivals make this sharper, because the audience is often watching a one-off performance. There is no second take. If the picture drops during the moment people tuned in for, the loss cannot be repaired afterwards.

• Router or switch restarts at the venue
• Upload bandwidth shared with other users on site
• Loose or damaged cabling near the encoder
• Short interruptions at the platform ingest point

Hot, warm and cold: how ready is the spare

Crews describe the readiness of a backup in three words. A hot backup is already sending the full programme to a second ingest address, so switching is a matter of moments. A warm backup is built and powered but not sending, and needs a short start-up. A cold backup is equipment on the shelf that can be set up if something goes wrong.

For a show with a live audience watching, we favour a hot path wherever the platform allows a second feed. The cost is extra upload capacity and a second encoder, but the gain is that the decision to switch is small and quick, rather than a rebuild under pressure.

Warm and cold options still have a place on small, low-risk streams. The point is to choose the level on purpose, agree it with the producer beforehand, and not discover on the day that everyone assumed a different one.

• Hot: sending now, switch in moments
• Warm: built and powered, short start-up
• Cold: held in reserve, set up on demand

Separate really means separate

The most common mistake is a backup that looks independent and is not. Two encoders plugged into the same network switch share a weak point. Two internet connections from one supplier may leave the building through the same cable. A second feed that uses the same ingest server fails with the first if that server has a problem.

Priya, who designs our delivery paths, checks this by tracing each route on paper from the encoder to the platform and asking one question at every step: what do these two routes share? Anything that appears in both lines is a single point of failure, and it is better to find it at a desk than during transmission.

Power matters too. If both encoders draw from one supply and a breaker trips, both paths go down together. Separate circuits, or a small battery unit on the encoders, are cheap compared with the show they protect.

• Different network connections, ideally different suppliers
• Different encoders on different power
• Different ingest addresses where the platform permits

Who decides to switch, and when

Having a backup is only useful if someone is watching for the fault and knows what to do. On our streams one person monitors the outgoing feed as a viewer would see it, on a separate screen fed from the platform, not from the gallery. The programme can look perfect in the control room while the audience sees a frozen frame.

The rule we agree before doors open is simple: if the monitored feed stalls for longer than a stated short period, the streaming lead switches to the backup without waiting for discussion, and tells the director over comms afterwards. Hesitation costs more than an unnecessary switch.

We also write down how to switch back. Returning to the main path is not urgent, so it happens at a quiet moment in the programme, once the fault has been understood and the main route has been steady for a while.

What to ask your streaming crew

If you are commissioning a live stream, you do not need technical vocabulary to check the arrangements. Ask whether there is a second route, and whether it shares anything with the first. Ask whether it is hot, warm or cold, and who is watching the stream as the audience sees it.

A good answer is specific: it names the connections, says how long a switch takes and says who makes the call. A vague answer, such as saying the equipment is reliable, tells you nothing about what happens when a single cable fails.

Raise these questions at the planning stage, when a second connection can still be arranged with the venue. On the day itself, the options have already been narrowed by what was brought and what was booked.

• Is there a second route, and what does it share with the first?
• Is it hot, warm or cold?
• Who watches the feed as viewers see it?