Removing SRT’s barriers to broadcast scale

By Jonathan White, Senior Director Support & TSS EMEA

Secure Reliable Transport (SRT) has earned its place in live contribution and distribution by solving a practical problem: moving high-quality video reliably across unpredictable networks. Its open foundations have supported broad adoption, while the economics and reach of the public internet offer a compelling alternative, or complement to satellite, dedicated fibre, and leased circuits.

But as more live content moves over SRT, the question changes. It is no longer simply whether SRT can carry a feed. It is whether it can support hundreds of services, high-bitrate UHD contribution, and mission-critical workflows without creating a sprawling server estate.

“Appear’s achievement is not changing what SRT is. It is removing the scale barriers that previously kept it from taking on the most demanding broadcast workflows.”

From CPU-bound to purpose-built

Appear’s first SRT implementation used the open-source LibSRT library. That approach worked, but its reliance on CPU processing limited the density we could achieve within the X Platform architecture. Therefore, we developed our own hardware-accelerated implementation. The CPU handles the SRT control layer, while the FPGA processes the media transport. Each part of the system does the work it is best suited to perform.

That separation changes the operating envelope. One IP gateway module can support up to 128 services and 6 Gb/s of SRT traffic, with retransmission bandwidth controlled at service level. Configured as a dedicated SRT gateway with 100Gb input connectivity, a 2RU X20 scales to 1,536 connections and 72 Gb/s of SRT traffic. At nominal service rates, that is equivalent to 72 UHD JPEG XS services at 1 Gb/s or around 900 UHD HEVC services at 80 Mb/s. In a combined HEVC encoding-and-SRT configuration, the platform can support up to 22 simultaneous UHD camera feeds.

The result is predictable capacity. Hardware processing reduces packet-processing overhead. End-to-end latency remains governed by the configured buffer and network conditions.

Density changes the economics

At small scale, software can be the right answer. Receiving one or two feeds in the cloud may not justify dedicated hardware, which is why VX uses Appear’s SRT implementation in software. At a remote venue, X5 combines compact AVC/HEVC processing with SRT transport. For latency-sensitive workflows, X5, X10 and X20 can also transport JPEG XS over SRT, combining low-latency compression with SRT’s resilience. X10 and X20 address workflows where density, modularity, and deterministic performance matter.

The distinction becomes clearer as operations grow. A server estate brings rack space, power, cooling, operating-system maintenance, patching, and additional points of failure. FPGA acceleration consolidates that workload into fewer systems and provides known capacity per module. When requirements increase, operators can add another module rather than another server.

This is a total-cost-of-ownership argument, not simply a hardware-performance argument.

Software SRT and hardware-accelerated SRT, at a glance

Software SRT (VX)

• Scale: a few feeds, cloud-based ingest
• Best suited to: small-scale reception where dedicated hardware isn’t justified
• Economics: no dedicated hardware investment needed at this scale

Hardware-accelerated SRT (X Platform: X5, X10, X20)

• Scale: up to 128 services and 6 Gb/s per IP gateway module; up to 1,536 connections and 72 Gb/s on a dedicated 2RU X20 gateway
• Best suited to: density, modularity, and deterministic performance at the edge and the core
• Economics: consolidates rack space, power, cooling, and patching into fewer systems — a total-cost-of-ownership argument, not just a performance one

Proven across different network realities

Recent deployments show how those characteristics translate into operations. In Chile, hardware-accelerated SRT carried an HEVC contribution feed from Chilevisión into AWS, supporting cloud processing for a nationwide 4K FIFA World Cup service.

World Archery now sends SRT feeds from venues worldwide to its production centre in Lausanne, before distributing coverage to more than 15 media partners. Bringing production in-house gave the federation greater control and a calculated 18-month payback period.

Corus Entertainment is using SRT as its primary transport protocol across Canadian affiliates, news bureaux, and operational hubs, replacing its reliance on leased circuits while retaining legacy interfaces where the transition requires them.

These are different workflows, budgets, and risk profiles. The common requirement is reliable transport without unnecessary operational complexity.

Different workflows, same requirement

Customer	Workflow	Result
Chilevisión	Hardware-accelerated SRT carrying an HEVC contribution feed into AWS for cloud processing	Nationwide 4K FIFA World Cup service
World Archery	SRT feeds from venues worldwide into a production centre in Lausanne, distributed to more than 15 media partners	Greater production control and an 18-month payback period
Corus Entertainment	SRT as primary transport across Canadian affiliates, news bureaux, and operational hubs	Replaced reliance on leased circuits, while retaining legacy interfaces where needed

SRT’s next chapter is scale

SRT will continue to evolve. Appear’s implementation supports bonding, rendezvous mode, and ST 2022-7 protection. Richer monitoring remains a natural area for development. Yet SRT’s central value remains straightforward: reliable, secure transport over networks that are available almost anywhere.

Hardware acceleration allows that value to extend from a few feeds at the edge to high-density contribution and distribution at the core. SRT already works. Appear makes it work at the scale live production now demands.