Secure Reliable Transport (SRT) is a protocol designed to address issues such as packet loss, jitter, delay and bandwidth issues. All can get in the way when streaming over unpredictable network conditions. SRT can deliver a stable stream, even when conditions are less than ideal.
SRT is most often used for first-mile contribution rather than last-mile delivery and is merely a wrapper around your video content—be it MPEG-2, H.264 or even HEVC, so it can used with any video and audio codec.
How does it work?
SRT uses its own packet loss recovery method over UDP. This saves resources by avoiding repeat transmission requests and will only request that the missing packets will be resent. For example, you’re sending an SRT video stream with packets 1 through 10, but on the receiving end at Wowza, the stream is missing packets 3 and 5. SRT will request only those two missing packets (3 and 5) as opposed to requesting all of 10 of them be resent.
The speed at which this takes place will allow you to deliver a near flawless stream to your viewer without delay or jitter. It is considered a low-latency protocol with a sub-3 second average.
When to use it:
• Remote event production on less-than-ideal networks
• Any streaming workflow complicated by first-mile challenges
• Multi-site streaming events delivered across the globe
• Broadcasters requiring low-latency content acquisition
Wowza SRT support
There are two ways to output an ingested SRT stream in Streaming Engine: either you can let Wowza Streaming Engine transmux the SRT stream into a supported playback format such as MPEG-DASH, HLS, RTMP, HDS, Microsoft Smooth Streaming, and RTSP/RTP, or you can deliver the stream as SRT by setting up a stream target. Your SRT destination might be almost anything—a third-party CDN or a video-sharing site, perhaps.
When ingesting the stream, Wowza Streaming Cloud functions in SRT listener mode when establishing a handshake. Currently, Wowza Streaming Cloud with SRT is ingest only, so the stream will need to be transmuxed for playback.
Haivision released SRT on GitHub in 2017 as an open-source technology stack and protocol. If you’re interested in learning more about this technology and what the plans are for SRT in the near future, please visit the SRT Alliance Page.