SRT Distribution: One SRT Feed, Unlimited Bit-Exact Copies

Ask ten broadcast engineers what "distribution" means and you will get ten slightly different answers, because the word gets stretched across a signal chain that is really two jobs stitched together. If you are researching SRT distribution — taking one live feed and getting it out to many places at once — it pays to be precise about where in that chain you actually are. So let us start there, then get to the part that costs people their evenings: fanning a single feed out to a growing list of downstream partners.

The broadcast signal chain: contribution vs distribution

A live signal travels in two broad stages, and SRT — Secure Reliable Transport, the UDP-based protocol with ARQ error recovery and AES encryption that Haivision open-sourced in 2017 — is used across both. Understanding which stage you are in tells you what problem you are actually solving.

Contribution: getting the feed in

Contribution is the upstream half. It is the job of moving a feed from the source — a stadium, an OB truck, a remote camera position, a correspondent in the field — to a hub: a studio, a playout centre, or a cloud ingest point. Historically this meant a satellite uplink or a booked fibre circuit. Contribution feeds tend to be high-bitrate, few in number, and mission-critical: if the contribution link drops, there is nothing to distribute.

Distribution: sending the finished feed out

Distribution is the downstream half. Once the finished signal is sitting at the hub, distribution is the job of sending it out to everyone who needs it — broadcasters, telcos, CDNs, affiliates, satellite operators. Where contribution is few-to-one, distribution is one-to-many. This is where the fan-out problem lives, and it is what most people mean when they search for SRT distribution.

The two stages are not symmetric, and the difference matters operationally. A contribution failure is a single point of failure you watch like a hawk. A distribution failure is death by a thousand cuts — any one of dozens of downstream links can flap, and each one is a different partner phoning to ask why their feed just went black.

The distribution problem, stated plainly

Here is the scenario SRT distribution has to solve. You have one clean live feed at your hub. Ten, twenty, fifty downstream destinations each need their own copy of it, delivered reliably, right now. Traditionally there were three ways to do that, and none of them is cheap or quick:

• Satellite. Book a transponder, and every dish pointed at it gets the feed. Genuinely one-to-many and genuinely reliable — and genuinely expensive, slow to arrange, and weather-dependent.
• Dedicated fibre or managed circuits. Rock-solid and low-latency, but every new destination is a provisioning project with a lead time and a monthly line item, and you are tied to the routes your carrier serves.
• Racks of appliances. Put a gateway or an encoder at each site, wire SRT between them, and manage the mesh yourself. Flexible, but now you own hardware, licences, and uptime at every endpoint.

SRT changed the economics of all three by making the ordinary public internet reliable enough for professional feeds. Because SRT recovers lost packets with ARQ inside a bounded latency window, and encrypts everything with AES, you can push a broadcast-grade signal across links you do not control and expect it to hold. For live stream distribution, that solved the transport problem outright.

What the protocol alone does not solve is the operational weight of doing this at scale by hand — the part vendors tend to gloss over. Every downstream destination is:

• a connection to configure — caller or listener, ports, passphrase, stream ID, one set per partner;
• a latency figure to tune — set as a multiple of that link's round-trip time, measured, then adjusted for observed packet loss;
• a session to monitor — bitrate, retransmits, dropped packets — and something to restart at 2 a.m. when the link flaps.

One or two of those is a pleasant afternoon. Forty of them is a full-time job, and it grows linearly: every new taker adds another connection to babysit. Hand-rolled SRT fan-out does not fall over because SRT is hard — it falls over because the operational load scales with your partner list and your team does not.

SRT Cloud: purpose-built SRT distribution

SRT Cloud exists to take that operational load off the table. The model is deliberately narrow, and the narrowness is the point.

You send one SRT input. SRT Cloud produces unlimited bit-exact, 1:1 copies of it and delivers each copy to a taker — a downstream destination such as a broadcaster, telco, affiliate, CDN, or satellite operator. There is no transcoding and no re-encoding. Every output is frame-for-frame identical to the feed you sent; what leaves your encoder is exactly what each taker receives, down to the bit. Real takers on the platform today include FreeSat, Pyur, Eutelsat, SES, Etisalat, Swisscom, UIG, and KiwiSat.

Because it is a managed cloud service, there is no hardware to rack, no media server to babysit, and no long-term commitment. You point one input in, and you add or remove outputs as partners come and go. Pricing is simply €99 per output per month — one licence equals one SRT output — after a free trial, and you scale up or down whenever you like. Your bill tracks your partner list, not your procurement forecast.

The old way vs SRT Cloud

When dedicated hardware still wins

Now the honest part, because pretending otherwise helps nobody. A managed copy-and-deliver service is not the right tool for every distribution job, and there are serious products that earn their place.

Zixi runs a large managed-network business built around a protocol and broker fabric for guaranteed delivery across long, complex routes. Net Insight comes from the transport-network world and is at home in tightly engineered media backbones with precise timing. Haivision — who created SRT — make battle-tested gateways and encoders that have survived years in OB trucks and master control rooms, with deep protocol conversion between SRT, RTP, RTMP and MPEG-TS. If your plant is on-premises and conversion-heavy — if the job is remuxing between container formats, transcoding to several bitrates, or bridging half a dozen transports before a feed leaves the building — that is gateway work, and those vendors do it well.

SRT Cloud does not try to compete on that ground. It competes on one job: simple, cheap fan-out of a single feed to many destinations. If you need conversion and on-prem control, buy the appliance — and let SRT Cloud handle the distribution leg alongside it. If what you actually have is one feed and a lengthening partner list, the simpler tool wins, and it wins on price and on how little of your week it costs you.

The short version

Broadcast contribution and distribution are two jobs: contribution gets the feed to your hub, distribution gets it out to the world, and SRT is used across both. For the distribution half — the one-to-many fan-out that used to mean satellite time, booked fibre, or a rack at every site — SRT Cloud turns it into one input, unlimited bit-exact copies, and a €99-per-output bill. One feed in, unlimited exact copies out.