Media stream

Description

A media stream is a core concept in the 3GPP FLUS (Framework for Live Uplink Streaming) architecture, defined in specifications 26.238 and 26.939. It refers to the content that is transmitted from a FLUS source to a FLUS sink within the context of a media session. While a media session defines the temporal and control boundaries for media delivery, a media stream represents the actual payload—the continuous flow of media data. This could include audio, video, or auxiliary data such as metadata or subtitles. A media session may contain one or more media streams, allowing for the multiplexing of different media components that are synchronized and delivered together.

The architecture of a media stream involves the encapsulation of media content into a format suitable for transmission over the network. This typically involves encoding the media using a codec, packetizing it into a transport protocol such as RTP (Real-time Transport Protocol) or HTTP, and then sending it over the user plane. The media stream is associated with specific parameters, such as codec type, bitrate, resolution, and frame rate, which are negotiated during the establishment of the media session. The FLUS source is responsible for generating the media stream, while the FLUS sink receives and processes it. The control protocol of the media session manages the setup and teardown of the media streams, as well as any dynamic adjustments.

In operation, a media stream is created when the FLUS source begins capturing and encoding media content. The stream is then transmitted to the FLUS sink, which may be a media server, a rendering device, or a processing entity. The media stream can be unidirectional (from source to sink) or bidirectional, depending on the application. For live uplink streaming, the stream is typically unidirectional, carrying content from a camera or microphone to a production studio or content delivery network. The media stream may be subject to QoS treatment, ensuring that it receives the necessary bandwidth and low latency for real-time delivery. Network slicing and edge computing can be leveraged to optimize the path and processing of the media stream.

The media stream concept is closely related to the media session, as a media stream cannot exist outside of a media session. The media session provides the control and management framework, while the media stream provides the actual content. This separation of concerns allows for flexible and scalable media delivery. For example, a single media session could carry multiple media streams representing different camera angles, and these streams could be individually controlled or switched. The media stream is also associated with synchronization information, such as timestamps, to ensure that multiple streams are played back in sync.

Overall, the media stream is the fundamental unit of media content delivery in FLUS. Its definition has remained consistent across Releases 15 to 19, with only a minor wording change in Release 16 to align with the updated media session definition. This stability reflects the maturity and clarity of the concept within the 3GPP framework.

Purpose & Motivation

The media stream concept exists to provide a clear abstraction for the actual media content being delivered in a FLUS session. It separates the content from the control and management functions, allowing each to be designed and optimized independently. This separation is crucial for flexibility: a media session can manage multiple media streams, each with different characteristics, without complicating the control plane. It also enables interoperability, as media streams can be encoded and transported using standardized protocols, ensuring that different vendors' equipment can work together.

Before the introduction of FLUS, live uplink streaming often treated media content as a monolithic flow, making it difficult to manage multiple components or adapt to changing conditions. The media stream concept addresses this by allowing individual streams to be identified, prioritized, and manipulated. For example, in a live broadcast, the video stream might be given higher priority than an auxiliary data stream. This granularity is essential for delivering high-quality, reliable live media over 5G networks.

Furthermore, the media stream concept supports advanced use cases such as multi-view streaming, where multiple video streams are delivered simultaneously and synchronized. It also facilitates dynamic adaptation, where the bitrate or resolution of a stream can be adjusted in real-time based on network conditions. This is particularly important for mobile live streaming, where bandwidth can fluctuate. By defining media streams as distinct entities, 3GPP enables a rich set of media services that can leverage the full capabilities of 5G, including ultra-reliable low-latency communication (URLLC) and enhanced mobile broadband (eMBB).