Menu
Latest
12 meetings · Prague, Jeju · 12–16 Oct
Reports
RAN1#126-bis · SA2#177

Glossary term · QoS

5G QoS Flow

5G Quality of Service Flow

QoS →

5G QoS Flow is the finest granularity of QoS forwarding treatment in the 5G System, where all traffic mapped to one flow receives identical treatment such as scheduling, queue management, rate shaping, and RLC configuration.

Introduced
Rel-15
Where
Core Network › 5G Core
Specifications
2 specs
5G QoS Flow Description Purpose What changed Neighbours Specifications

Description

The 5G QoS Flow is the most granular QoS treatment unit in the 5G System (5GS). It is defined in 3GPP TS 23.501 as the finest granularity for QoS forwarding treatment, meaning that every packet mapped to a given QoS Flow receives the same treatment in the network — the same scheduling policy, queue management policy, rate shaping policy, RLC configuration and so on. If two traffic streams require different forwarding treatment, they must be carried on separate 5G QoS Flows. This concept replaces the EPS bearer as the primary QoS differentiation mechanism used in 4G, while still supporting a bearer-like model through the QoS Flow Identifier (QFI).

Architecturally, QoS Flows sit between the PDU Session and the actual data radio bearers. A PDU Session may contain multiple QoS Flows, each identified by a QFI that is carried in the encapsulation header of the user plane (e.g. GTP-U extension header or SDAP header). The Session Management Function (SMF) is the control-plane authority that creates, modifies and deletes QoS Flows, using policy information received from the Policy Control Function (PCF) via the N7 reference point and the Npcf_SMPolicyControl service. The SMF then instructs the User Plane Function (UPF) over N4 (PFCP) and the Access and Mobility Management Function (AMF) over N11, which in turn signals the RAN over N2. The RAN (gNB) maps QoS Flows to Data Radio Bearers (DRBs) based on the QoS parameters and the NG-RAN's own radio resource management decisions.

Each QoS Flow is characterised by a 5G QoS Identifier (5QI) that references a set of standardised QoS characteristics (resource type, priority level, packet delay budget, packet error rate, averaging window, maximum data burst volume). QoS Flows are further classified as Guaranteed Bit Rate (GBR), Delay-Critical GBR, or Non-GBR. GBR flows carry additional QoS parameters such as Guaranteed Flow Bit Rate (GFBR) and Maximum Flow Bit Rate (MFBR), which the RAN and UPF enforce through admission control and rate enforcement. Non-GBR flows rely on the 5QI characteristics alone. In addition, a Reflective QoS mechanism allows the UE to derive uplink QoS rules from downlink packets, reducing signalling overhead.

The mapping between application traffic and QoS Flows is governed by QoS rules. In the downlink, the UPF performs packet detection and classification using Packet Detection Rules (PDRs) received from the SMF, and marks packets with the appropriate QFI. In the uplink, the UE applies QoS rules (explicitly signalled or reflectively derived) to map traffic to QoS Flows, and the SDAP layer in the UE and gNB handles the mapping between QoS Flows and DRBs. The SDAP (Service Data Adaptation Protocol) is the key new protocol introduced in 5G specifically to decouple QoS Flow-level treatment from radio bearer-level treatment, giving the RAN flexibility to multiplex multiple QoS Flows onto a single DRB when their QoS requirements are compatible.

The 5G QoS Flow model is central to network slicing, because each slice can be associated with its own set of QoS Flows and 5QI profiles. It also underpins time-sensitive and industrial use cases through Delay-Critical GBR flows, and supports differentiated treatment for enhanced Mobile Broadband (eMBB), Ultra-Reliable Low-Latency Communication (URLLC) and massive Machine-Type Communication (mMTC). Because the definition has remained stable across Rel-15 through Rel-20, the concept has proven robust and is now the foundational QoS abstraction for all 5G services.

Purpose & Motivation

The 5G QoS Flow exists to provide a scalable, flexible and fine-grained QoS differentiation mechanism that overcomes the limitations of the 4G EPS bearer model. In LTE, QoS was tied to bearers — EPS bearers and E-RABs — which created a rigid one-to-one coupling between QoS treatment and radio bearer configuration. Establishing a new bearer required significant control-plane signalling (dedicated bearer activation procedures), and the number of bearers per UE was limited. This made it difficult to support the diverse mix of services envisioned for 5G, from low-latency industrial control to high-throughput video, on the same UE and PDU Session.

The QoS Flow concept decouples QoS treatment from radio bearer realisation. Multiple QoS Flows can be multiplexed onto a single DRB, and the RAN can decide how to map flows to bearers based on radio conditions and scheduling policy. This reduces signalling overhead and gives the radio access network more freedom in resource management. It also allows the core network to apply per-flow policy without forcing the RAN to instantiate a dedicated bearer for every service.

Another motivation was the need to support network slicing and policy-driven QoS. With QoS Flows, the PCF can define per-application policies that the SMF translates into flow-level rules, and the UPF enforces them in the user plane. This aligns with the service-based architecture of the 5G core and enables operators to offer differentiated SLAs per slice, per application or per subscriber. The introduction of Reflective QoS further reduces signalling by letting the UE mirror downlink treatment in the uplink, which is particularly valuable for IoT devices and high-density deployments.

What changed in its specifications

The latest release-to-release comparison of specifications that define 5G QoS Flow. Each summary describes the whole specification, not only this term.

TS 23.501Rel-18 → Rel-19

Rel-19 added MPQUIC steering to multi-access PDU sessions, XR and media handling, and an Energy Information Function. It also added Local Offloading Management for edge computing, Vertical Federated Learning for AI/ML, and Ambient IoT text.

Full comparison in TS 23.501 →
TS 29.512Rel-18 → Rel-19

Rel-19 adds multi-access ATSSS capabilities, QoS monitoring with available-bitrate parameters, and XR/media policy provisioning such as data burst size and expedited transfer. It also adds local offloading management, non-3GPP device information, and URSP provisioning in EPS support.

Full comparison in TS 29.512 →

Release Timeline

Evolution Across Releases

An overview written from the defining specifications, not a verified change history. For the exact text, open the specifications listed below.

Rel-15 Initial

Initial definition introduced the 5G QoS Flow as the finest granularity for QoS forwarding treatment in the 5G System, with all traffic mapped to the same flow receiving identical treatment (scheduling, queue management, rate shaping, RLC configuration). Providing different treatment requires separate flows.

The definition text carried over unchanged from Rel-15; the concept remained stable while supporting features such as URLLC enhancements and reflective QoS were refined in related specifications.

The definition text carried over unchanged from Rel-15; Rel-18 work on 5G-Advanced extended the surrounding QoS framework (e.g. for XR and media) without altering the core definition.

The definition text carried over unchanged from Rel-15; only formatting/line-wrapping differs in the quoted text, with no substantive change to the meaning.

The definition text carried over unchanged from Rel-15; the term remains the foundational QoS abstraction for 5G-Advanced and future releases.

Defining Specifications

3GPP specifications that define or reference 5G QoS Flow, with the latest known release. Sourced from the 3GPP document catalog — see methodology.

SpecificationTitleRelease
TS 23.501 vk30 5G System Architecture Rel-20 read
TS 29.512 vk10 Session Management Policy Control Service Stage 3 Rel-20 read