Connection mode

Description

Connection mode is a fundamental concept in telecommunications that describes the type of association between two communicating entities, as required by the bearer service for the transfer of information. It is defined in 3GPP specifications 21.905 (Vocabulary) and 22.100 (UMTS core network architecture). The term distinguishes between two modes: connection-oriented and connectionless. In connection-oriented mode, a logical association called a connection must be established between the source and destination entities before any information can be exchanged. This connection has a lifetime that spans from establishment to release. In connectionless mode, no prior connection is established; each message must carry the source and destination network addresses, and there is no guarantee of ordered delivery. The lifetime of a connectionless bearer service is reduced to the transport of a single message.

The concept of connection mode is rooted in the design of transport networks and has been adopted in 3GPP specifications to describe bearer service characteristics. In a connection-oriented bearer service, the network establishes a path or context between the endpoints, which may involve resource reservation, state maintenance, and signaling. This mode is suitable for applications that require reliable, ordered delivery, such as voice calls, video streaming, and file transfers. The connection setup may involve procedures like PDP context activation in GPRS/UMTS or PDU session establishment in 5G. Once established, the connection provides a virtual circuit over which data packets are forwarded with minimal per-packet overhead.

In contrast, connectionless bearer services treat each message independently. There is no connection setup or teardown; the sender simply transmits a message with the destination address, and the network routes it based on that address. This mode is efficient for short, bursty messages, such as DNS queries, SMS, or IoT sensor readings. However, it offers no guarantee of delivery order, and packets may be lost or duplicated. The connectionless mode is often used in IP networks, where the Internet Protocol itself is connectionless, but higher-layer protocols like TCP provide connection-oriented semantics.

In 3GPP networks, the connection mode of a bearer service is determined by the QoS requirements and the nature of the application. For example, conversational voice and streaming video typically use connection-oriented bearers to ensure consistent quality, while background traffic or signaling may use connectionless bearers. The core network elements, such as the SGSN, GGSN, and later the MME, SGW, and PGW, manage the establishment and release of connections. In 5G, the Session Management Function (SMF) handles PDU session establishment, which is a connection-oriented concept, while some signaling may be connectionless.

The definition of connection mode has remained unchanged across all releases from Rel-4 to Rel-19, with only minor editorial formatting changes. This stability reflects its foundational nature in telecommunications. However, the implementation of connection-oriented and connectionless services has evolved with the introduction of new technologies like GPRS, UMTS, LTE, and 5G, each with its own bearer and session management mechanisms. The term continues to be relevant as networks support diverse applications with varying QoS requirements.

Purpose & Motivation

Connection mode exists to categorize bearer services based on their need for a pre-established logical association. In the early days of telecommunications, networks were either circuit-switched (connection-oriented) or packet-switched (connectionless). As networks evolved to support both voice and data, the need arose to define bearer services that could operate in either mode depending on the application's requirements. The connection mode concept provides a clear framework for specifying how data is transferred between endpoints.

The primary problem it solves is the need for different levels of reliability, ordering, and overhead. Connection-oriented services provide reliable, ordered delivery with a setup phase, which is essential for real-time and high-integrity applications. Connectionless services provide low overhead and are suitable for short, independent messages. By defining these modes, 3GPP allows network operators and application developers to choose the appropriate bearer service for each type of traffic, optimizing resource usage and performance.

Historically, the concept was formalized in the UMTS specifications (22.100) and the vocabulary (21.905) in Release 4. It built upon the OSI reference model's distinction between connection-oriented and connectionless network services. In 3GPP, the connection mode is a characteristic of a bearer service, which is a service that provides the capability to transfer information between two points. The definition has remained stable, serving as a common vocabulary term across all generations of mobile networks. Its importance is evident in the design of core network protocols, such as the GPRS Tunnelling Protocol (GTP), which can operate in both connection-oriented and connectionless modes depending on the context.