Skip to main content
Tech

Towards 6G Media

5G-MAG's early media requirements input to 6G standardisation, backed by an AI-driven testbed.

6G (IMT-2030) is the next generation of mobile standards, currently in the research and requirements phase at 3GPP and ITU-R (International Telecommunication Union Radiocommunication sector), with initial specifications expected from 2028. 5G-MAG actively contributes media-specific requirements and use cases to the 6G standardisation process, covering media delivery at extreme data rates, Artificial Intelligence (AI)-native network management for media traffic, immersive and haptic experiences, and the evolution of broadcast and multicast. For acronyms used here, see the Glossary.

In parallel, the 5G-MAG 6G Testbed provides an early experimental platform for AI-driven media traffic classification and optimisation over 5G networks, informing future 6G design. The testbed produces labelled traffic datasets and classification results that 5G-MAG uses to shape its requirement inputs to 3GPP.

Key standardisation bodies: ITU-R IMT-2030 (framework), 3GPP working group SA1 (requirements), 3GPP working groups SA2 and SA4 (architecture and media aspects). No normative 6G specifications exist yet.

Reference tools: An experimental testbed is available on the developer portal under 6G Testbed & AI Traffic.

Go deeper

Technology & Analysis

The IMT-2030 usage scenarios, standardisation timeline, and why an experimental testbed matters at the study stage.

Standards Tracking

5G-MAG's standards-tracking view of 6G study items and media requirements.

Software Tools

An experimental testbed for AI-driven media traffic classification and optimisation.

Execution Plan


Technical Documentation

The following resources are available:

Inputs from 5G-MAG to 3GPP Workshops on 6G

Each link below downloads a ZIP contribution package (the submitted document and its supporting files) from the 3GPP file server.

The IMT-2030 usage scenarios and where media sits

The global 6G framework is Recommendation ITU-R M.2160-0 (November 2023). It defines six usage scenarios. The table below maps them to their 5G (IMT-2020) origin and to media relevance.

IMT-2030 usage scenarioOriginMedia relevance
Immersive communicationEvolves eMBBPrimary scenario for high-rate and immersive media (volumetric, XR, high-resolution)
Massive communicationEvolves mMTCSensor and device data; indirect (metadata, telemetry)
Hyper-reliable and low-latency communication (HRLLC)Evolves URLLCInteractive and real-time media, remote production, haptics
Ubiquitous connectivityNewBroadcast/multicast reach, non-terrestrial delivery
AI and communicationNewAI-native traffic management for media; AI/ML model and inference traffic as a workload
Integrated sensing and communication (ISAC)NewContext capture for media applications; not a delivery path in itself

M.2160 also states overarching design principles (sustainability, security and resilience, connecting the unconnected, and ubiquitous intelligence) and extends the 5G capability set with new capabilities. These are targets, not specifications: the numeric values and the mechanisms to reach them are the subject of the 3GPP work below.

Standardisation timeline and where the work is

6G in 3GPP is a Release-21 target. The current activity is study-stage in Release 20.

  • ITU-R completed the framework (M.2160) in 2023 and is progressing the detailed technical performance requirements and evaluation methodology for IMT-2030 through Working Party 5D.
  • 3GPP SA1 produced TR 22.870 (Study on 6G Use Cases and Service Requirements) at the study stage; the normative Stage 1 6G requirements are expected in Release 21.
  • 3GPP SA2 and SA4 are running architecture and media studies. SA4's media study is TR 26.870.
  • Initial 6G specifications are expected around 2028.

No normative 6G specification exists yet. Everything on this page that concerns 6G behaviour is study-stage and provisional. Where a number or requirement is not yet decided in 3GPP or ITU-R, it should be treated as an open question rather than a settled value.

Study items relevant to media in 6G

The following studies are the current primary sources. TR 22.870 and TR 26.870 are the 6G studies; the AI/ML studies are 5G Advanced work that the 6G media work builds on.

StudyGroupTrackFocus
TR 22.870SA16G Stage 16G use cases and candidate service requirements, clustered around the IMT-2030 usage scenarios
TR 26.870SA46G mediaMedia aspects of a 6G system: media services, formats, traffic and delivery
TR 26.847SA45G media AI/MLEvaluation of AI and ML in 5G media services; evaluation testbeds, scenarios and metrics
TR 22.874SA15G AI/MLTraffic characteristics and performance requirements for AI/ML model transfer in 5GS

TR 26.847 and TR 22.874 matter for 6G because two of the new IMT-2030 scenarios (immersive communication and AI and communication) drive traffic that these studies already characterise: high-rate media flows on one side, and model transfer flows on the other. See the AI & ML standards page for the interface and procedure detail.

Why an experimental testbed at the study stage

Standardisation at the study stage is requirement-driven, and requirements are stronger when backed by measurement. AI workloads (large language models, image and video generation, agentic multi-step tool calling, real-time multimodal interaction) produce traffic patterns that differ from classic media streaming. These include bursty request/response cycles, asymmetric uplink/downlink ratios, long-lived sessions with variable think-time, and sensitivity to time-to-first-token as well as to sustained throughput. The 5G-MAG 6G Testbed exists to characterise these patterns under controlled, reproducible network conditions.

Concretely, the testbed:

  • Emulates realistic network conditions on a single machine using Linux Traffic Control (tc) with netem and Hierarchical Token Bucket (HTB), including profiles derived from 3GPP 5QI values (see TS 23.501, the standardised 5QI to QoS-characteristics mapping) and candidate 6G conditions (for example a hyper-reliable low-latency profile).
  • Runs AI and media workloads over those conditions and captures traffic at L3/L4 (via tcpdump) and optionally L7 (via mitmproxy).
  • Logs metrics for large-scale, reproducible analysis, producing labelled datasets and classification results.

These outputs are what 5G-MAG uses to ground its media and AI-traffic requirement inputs to 3GPP. Full detail, including the emulator YAML profile schema and the AI characterisation architecture, is on the 6G Testbed developer scope page.

References to verify

These identifiers on this page were not confirmed against a primary source (the 3GPP/ETSI portals block automated access): TR 26.870 (number and title "Study on Media Aspects for 6G System"). Verify against the 3GPP work plan before publication.

note

Refer to the Tech repository to contribute to this documentation.