Quality of Service Session

Description

A QoS session in 3GPP terminology represents the lifetime of a Packet Data Protocol (PDP) context — the period between the opening and closing of a network connection whose characteristics are governed by a specific QoS profile. This concept was introduced in Release 4 as part of the effort to formalize end-to-end quality of service differentiation in packet-switched networks. Each QoS session is associated with a distinct set of QoS parameters, including traffic class, guaranteed bit rate, maximum bit rate, transfer delay, and allocation/retention priority, as defined in 3GPP TS 23.107.

The architecture supporting QoS sessions relies on the bearer concept. In UMTS Release 99 and later, a PDP context activates a primary bearer, and additional secondary PDP contexts can be established to create multiple QoS sessions. Each session maps to a radio access bearer (RAB) and a core network bearer, forming an end-to-end tunnel with negotiated QoS attributes. The GGSN (Gateway GPRS Support Node) and SGSN (Serving GPRS Support Node) enforce the QoS profile through traffic policing, scheduling, and resource reservation mechanisms.

When a UE initiates a session, it sends an Activate PDP Context Request containing a requested QoS profile. The network may negotiate this profile downward based on available resources, subscription constraints, and policy rules from the PCRF (Policy and Charging Rules Function). The resulting QoS session is maintained until the PDP context is deactivated, either by the UE, the network, or due to timeout. During the session, the QoS profile can be modified through a PDP Context Modification procedure.

Multiple QoS sessions allow a single UE to simultaneously handle different types of traffic — for example, a conversational voice session alongside a background file download — each with its own guaranteed and best-effort characteristics. This differentiation is fundamental to supporting services such as VoIP, video streaming, and interactive gaming over cellular networks.

With the evolution toward 5G, the QoS session concept has been superseded by PDU sessions and QoS flows as defined in TS 23.501. However, the foundational principles — per-session QoS profiles, multiple concurrent sessions, and network-controlled QoS negotiation — remain central to 5G QoS architecture. The term itself is maintained in TR 21.905 for backward compatibility and reference purposes.

Purpose & Motivation

The QoS session concept was created to solve a fundamental problem in early packet-switched cellular networks: how to provide differentiated treatment to different traffic types within a single user's data connection. Before Release 4, GPRS networks offered only best-effort delivery with limited QoS differentiation. The introduction of QoS sessions allowed operators to guarantee specific bandwidth, delay, and reliability characteristics for individual data streams.

This was motivated by the growing demand for real-time services over cellular networks. Voice over IP, video telephony, and streaming media required performance guarantees that best-effort delivery could not provide. By defining QoS sessions with distinct profiles, 3GPP enabled the network to allocate resources appropriately and enforce service level agreements.

The concept also addressed the need for multiplexing multiple services over a single PDP address. A user could maintain a conversational session for voice while simultaneously running a background session for email or web browsing, each receiving appropriate QoS treatment. This multi-session capability was a significant advancement over the single-context model of earlier releases.