Retrieval service

Description

A retrieval service is an interactive telecommunications service that provides the capability of accessing information stored in database centres. The information is sent to the user on demand only, meaning that no information is delivered until the user explicitly requests it. Furthermore, the information is retrieved on an individual basis: the time at which an information sequence is to start is under the control of the user. This definition is sourced from ITU-T Recommendation I.113, which provides the vocabulary for integrated services digital networks (ISDN) and broadband ISDN, and it has been adopted into 3GPP terminology via specifications such as TS 21.905 (Vocabulary for 3GPP Specifications) and TS 22.060 (General Packet Radio Service (GPRS); Service description; Stage 1).

Architecturally, a retrieval service relies on a client-server model. The user's terminal equipment (e.g., a mobile station, UE, or computer) acts as the client, while database centres store the information content. The network provides the transport bearer that connects the client to the server. In a 3GPP context, this transport is typically a packet-switched connection such as GPRS, EPS, or 5GS. The retrieval service itself is an application-layer service that sits above the bearer services. It does not define the specific protocols for database access; rather, it characterizes the interaction pattern: user-initiated, on-demand, and individually controlled.

The key operational characteristic is that the user controls the start of each information sequence. This distinguishes retrieval services from distribution services (e.g., broadcast or multicast), where the network or content provider determines when information is sent. In a retrieval service, the user might browse a menu, select a specific item, and then receive that item. Examples include web browsing, file transfer, database queries, and video-on-demand. The service is interactive because the user can send requests and receive responses, potentially in a conversational manner.

In 3GPP specifications, the term appears primarily in vocabulary and service description documents. It is not a specific protocol or network function but a service category used for classification. TS 22.060, for instance, describes GPRS services and may reference retrieval services as one type of application that GPRS can support. The definition has remained stable across releases from Rel-4 through Rel-19, indicating its foundational nature. The only variations in the quoted definitions are formatting differences (line breaks and spacing) rather than substantive changes.

For telecom engineers and students, understanding retrieval services is important for classifying applications and mapping them to appropriate bearer capabilities. For example, a retrieval service may require low latency for interactive response, high throughput for large file downloads, or asymmetric bandwidth (more downlink than uplink). These requirements influence the choice of QoS parameters and network configuration. The concept also helps in distinguishing between different traffic types when designing networks and charging mechanisms.

Purpose & Motivation

The retrieval service concept exists to provide a standardized classification for interactive applications where the user initiates and controls the flow of information. It originated in ITU-T I.113 as part of the vocabulary for ISDN and broadband ISDN, aiming to categorize teleservices and bearer services. By defining retrieval services separately from distribution services (e.g., broadcast), it enables consistent descriptions of user interaction patterns and helps network designers allocate appropriate resources.

In the 3GPP context, adopting this term allows specifications to refer to a well-understood service category when describing capabilities of packet-switched networks like GPRS, EPS, and 5GS. It solves the problem of ambiguous terminology: without a common definition, different stakeholders might interpret "interactive service" differently. The retrieval service definition clarifies that the user controls the start time and that information is sent on demand, which has implications for traffic modeling, QoS, and charging.

The concept also addresses the need to differentiate between services that are symmetric versus asymmetric, real-time versus non-real-time, and user-controlled versus network-controlled. This differentiation is crucial for designing networks that can support a wide range of applications, from web browsing to video streaming, each with distinct requirements. Although the term itself is not a technical mechanism, it provides a shared vocabulary that underpins service descriptions and requirements in 3GPP specifications.