Your browser is not supported. Please update it.

10 November 2022

RIST dreams of mainstream with Source Adaptation feature

The Reliable Internet Stream Transport (RIST) protocol has received the first of what promises to be a slew of feature updates to further improve reliability. It is a timely reminder of the flexibility of RIST over the more widely adopted SRT (Secure Reliable Transport) for delivery of live video over IP, with SRT limited by its core set of libraries.

The RIST Forum hopes the specification updates will springboard the protocol into more mainstream adoption, outside of its current loyal user camps in professional media workflows from news and sports contribution, remote production, affiliate distribution and primary distribution.

Called Source Adaptation, the latest enhancement to RIST specifications is designed to react to a situation where packet loss recovery – one of RIST’s trademark functions – is not possible, if network capacity falls below bitrate. To keep a stream healthy against this backdrop of unpredictable network environments, RIST’s Source Adaptation feature enables a stream receiver to provide feedback information to the source, to take actions such as reducing bitrate or re-routing the stream where possible.

Source Adaptation enables the source itself to react to unfavorable network conditions in this way, as an emergency back-up feature of sorts. Being open source, the actual algorithms are left at the discretion of the implementer – which can both benefit the community with technical comments submitted, and potentially lead to patent infringement claims.

It appears Source Adaptation is adapted for cellular connections, where bandwidth fluctuations are common. If available bandwidth falls below stream bitrate, then no amount of ARQ (Automatic Repeat Request) – the packet loss recovery scheme – will recover lost packets.

This makes the RIST specification more like HLS and DASH protocols, which can reduce bitrate or re-route traffic in a receiver-driven fashion. Senders provide a number of discrete bitrates, and the client picks the one most suitable to current network conditions. But HLS and DASH alone do not consider latency (hence the development of low latency versions LL-HLS and LL-DASH) in one-to-many distribution scenarios.

Latency is important in one-to-one transmission scenarios, where the sender can control transmission rate as a function of network conditions. But until now, there was no recognized interoperable standard or specification for enabling encoders to switch bitrate on-the-fly seamlessly. RIST Source Adaptation aims to address this.

The diagram below shows a simple example of Source Adaptation, taken from the technical recommendations, which have been released by the Video Services Forum (VSF) in full, as TR-06-4 Part 1.

TR-06-01 details the first of RIST’s three specifications, RIST Simple Profile, published in 2018 and updated in 2021. TR-06-02 is the RIST Main Profile, published in 2021 and 2022, while TR-06-3 is the RIST Advanced Profile, published in 2021 and updated in 2022.

The VSF describes two aspects to Source Adaptation as extensions to RIST Simple Profile and RIST Advanced Profile. The first is that some sources, such as encoders with variable bitrate capability, can dynamically adapt their output based on network conditions. The second is that some sources using multiple network conditions in parallel can dynamically change the traffic mix over the various connections based on network conditions.

The RIST Activity Group anticipates the new Source Adaptation ancillary feature making a real difference for contributing content over the public internet.

Contributors to Source Adaptation include individuals from CBS/Paramount, Cobalt Digital, Arqiva, QVidium, Nevion, AlvaLinks, SupRadius/Ammux, Evertz, Zixi, Telecom Product Consulting, and Media Transport Solutions.

The latest update builds on RIST’s leap into the distribution game around the time of IBC 2022, with updates claiming to enable users to create a completely open-source alternative to proprietary DRM systems, for use in a multicast distribution deployment.

Distribution in this context likely refers to an internal project, or perhaps a low-volume B2B implementation, rather than a mass-market B2C implementation. Still, the door is open, and while RIST is currently facing off against SRT in the contribution market, success in distribution would likely accelerate adoption in contribution.

However, there is a sense that idealism has prevailed in the anti-DRM push from RIST, and that the cold harsh reality of the situation that DRM vendors have to battle daily on the device front could quickly bring it down to earth.

RIST has been found particularly effective for live sports use cases due to its support for higher bitrates than alternatives such as SRT.