FLUS

Framework for Live Uplink Streaming

Services →
Introduced in Rel-15 Also in: Core Network

FLUS is a 3GPP service enabler that standardizes the uplink of live media from a user's device to a network application, providing mechanisms for session control, ingestion, and quality management.

Category
Services
Introduced
Rel-15
Where
Services › Codecs
Also touches
1 segments
Specifications
13 specs
FLUS Description Purpose Related Classification Detected Changes Specifications

Description

The Framework for Live Uplink Streaming (FLUS) is a comprehensive set of 3GPP specifications designed to facilitate efficient, reliable, and high-quality live media streaming from User Equipment (UE) to a remote media server. It defines a client-server architecture where the UE acts as the media source (FLUS Client) and a network-based application function acts as the media sink (FLUS Server). The framework encompasses protocols and procedures for session negotiation, media transmission adaptation, and network exposure to optimize the streaming experience over potentially variable mobile radio conditions.

Architecturally, FLUS interacts with several 5G System components. The FLUS Client resides on the UE, while the FLUS Server is typically part of an Application Function (AF) in the 5G Core. The framework leverages the Network Exposure Function (NEF) or directly interfaces with the Policy Control Function (PCF) to request specific QoS treatment for the uplink media flow. This allows the FLUS service to request the establishment of a dedicated QoS Flow with guaranteed bit rate, priority, and packet loss characteristics suitable for real-time video encoding and transmission. Key procedures include capability discovery, where the UE and server exchange supported media codecs and FLUS features; session establishment, where streaming parameters are negotiated; and media control, which may involve commands from the server to the client to adjust bitrate, resolution, or other encoding parameters based on network feedback.

FLUS operates by establishing a control plane connection (e.g., over HTTP/2) between the FLUS Client and Server for session management. The actual media is then transported over a user plane data flow, typically using RTP over UDP/IP. A core innovation of FLUS is its integration with 5G network capabilities. The FLUS Server can invoke 5G network APIs to obtain analytics about the UE's radio conditions (e.g., expected throughput, latency) and can subsequently influence the network's treatment of the media flow through QoS policy control. Furthermore, FLUS defines media ingestion protocols that are efficient for live streaming, supporting mechanisms like timed metadata insertion, synchronization, and redundancy to improve robustness against packet loss. This end-to-edge framework ensures the live video feed is optimized from the camera sensor through the radio access network to the content delivery network, enabling broadcast-quality live streaming from mobile devices.

Purpose & Motivation

FLUS was created to address the growing market demand for high-quality, reliable live video streaming from mobile devices, a service popularized by social media platforms and content creators. Prior to FLUS, live streaming apps used proprietary, over-the-top (OTT) solutions that had no integration with the mobile network. These solutions struggled with unpredictable mobile network conditions, leading to poor video quality, buffering, and dropped streams. They could not leverage network intelligence or request prioritized resources, resulting in a subpar user experience, especially in crowded areas or at cell edges.

The primary problem FLUS solves is the lack of standardization and network awareness for uplink streaming. By creating a standardized framework, it enables interoperability between devices, networks, and application servers from different vendors. More importantly, it allows the streaming application to communicate its requirements to the 5G network, enabling network-assisted quality optimization. This addresses key limitations: 1) Inefficient reactive adaptation where the app only reacts to observed congestion, often too late. 2) Inability to guarantee resources, leading to contention with other background traffic. 3) Lack of radio analytics to proactively adjust encoding parameters. FLUS provides the hooks for the network to assist the application, paving the way for deterministic quality live broadcasting, which is critical for professional use cases like newsgathering, live sports production, and public safety communications.

Classification

Part of5GS
Related approachesAF

Release Timeline

Detected Changes Across Releases

from 3GPP Change Requests

Specific changes extracted from the „Change history“ tables of 3GPP specifications (8 CRs across 3 releases). Complements the general historical overview above with the evidence-based evolution of this function.

Studied in Rel-15, normative work from Rel-16.

Rel-16 6 changes

In Release 16, the Framework for Live Uplink Streaming (FLUS) was introduced to standardize uplink video streaming from devices like V2X UEs to an application server. This framework, involving a FLUS source (VAE client) and a FLUS sink (VAE server), established support within the Rx interface and the Npcf_PolicyAuthorization service for policy control. It defined a session-based procedure over the V1 reference point, capable of carrying streaming video over protocols like HTTP or RTP.

  • E-FLUS Updates TS 26.939CR0002
  • New Use Case and Correction for E_FLUS TS 26.939CR0007
  • Support of Framework for Live Uplink Streaming (FLUS) in Rx interface TS 29.214CR1632
  • Support of Framework for Live Uplink Streaming (FLUS) in Npcf_PolicyAuthorization service TS 29.514CR0162
  • Support of FLUS feature TS 29.514CR0208
  • Attaching back the file - 3GPP FLUS MO.xml TS 26.238
Rel-17 1 change

In Release 17, the FLUS (Framework for Live Uplink Streaming) function was enhanced by extending support for network-based media processing. This built upon the existing framework where a VAE client acts as a FLUS source and a VAE server acts as a FLUS sink to manage uplink video streaming sessions over the V1 reference point. The extension specifically allows the VAE server to receive and handle multiple videos from multiple FLUS sources for subsequent post-processing operations.

  • CR on the extending support for network-based media processing in FLUS TS 26.238CR0026
Rel-18 1 change

In Release 18, the FLUS (Framework for Live Uplink Streaming) function was removed from the 5GMSA_Ph2 specifications, and its associated workflows were updated accordingly. This change streamlined the architecture by deprecating the framework where the VAE client and VAE server acted as FLUS source and sink enablers for uplink video streaming.

  • [5GMSA_Ph2] Uplink streaming: removing FLUS and updating the workflows TS 26.501CR0042

Explore further

Broader topics and technologies where FLUS plays a role.

Defining Specifications

3GPP specifications that define or reference FLUS, with the latest known release. Sourced from the 3GPP document catalog — see methodology.

SpecificationTitleRelease
TS 23.795 vg10 V2X Application Architecture Study Rel-16
TS 26.238 vj00 Framework for Live Uplink Streaming (FLUS) Rel-19
TS 26.501 vj30 5G Media Streaming (5GMS) Architecture Rel-19
TS 26.804 vj10 5G Media Streaming Extensions Study Rel-19
TR 26.806 vi00 Technical Report on Smartly Tethering AR Glasses Rel-18
TR 26.857 vi00 Technical Report on Media Service Enablers Rel-18
TR 26.862 vh00 Immersive Teleconferencing & Telepresence for Remote Terminals Rel-17
TS 26.891 vg00 Media Distribution Services in 5G System Rel-16
TR 26.928 vj00 Study on eXtended Reality (XR) in 5G Rel-19
TR 26.939 vj00 Framework for Live Uplink Streaming (FLUS) Rel-19
TR 26.998 vj00 5G AR/MR Glasses Integration Study Rel-19
TS 29.214 vj20 Policy and Charging Control over Rx Rel-19
TS 29.514 vj40 5G System; Policy Authorization Service; Stage 3 Rel-19
Patrick Zandl

About the author: Patrick Zandl (b. 1974)

Telecommunications specialist, technology journalist (founder of the Mobil server), and developer who has been running since 2025 — the largest Czech-language resource on AI-assisted programming. Formerly Chief Wizard Architect at Prusa3D and head of development for Turris at CZ.NIC; currently a consultant and instructor on AI implementation in companies.