Service Provider Identification

Description

Service Provider Identification (SPI) is a parameter within 3GPP systems that unambiguously identifies the provider of a specific service or set of services to a user. It is distinct from subscriber identifiers like IMSI or MSISDN. The SPI is used by network functions to determine which policies, charging rules, and quality-of-service treatments to apply for a given data session or service flow. It is a key element in enabling service-aware networks.

Architecturally, the SPI is often configured in the user's subscription profile stored in the HSS (for packet core) or HLR. During session establishment, such as a PDN connection in EPS or a PDU session in 5GS, the relevant network node (e.g., MME, SMF) retrieves the SPI along with other subscription data. This SPI is then provided to policy and charging control functions like the PCRF (in 4G) or PCF (in 5G). The PCRF/PCF uses the SPI, in combination with other parameters like the APN or DNN, to select the appropriate Policy and Charging Control (PCC) rules. These rules dictate aspects like bandwidth limits, QoS class identifiers (QCIs/5QIs), and charging methods specific to that service provider's offering.

How it works in practice: An MVNO (Mobile Virtual Network Operator) leasing capacity from an MNO (Mobile Network Operator) will have its own unique SPI. When an MVNO subscriber attaches to the network, the MNO's policy system identifies the SPI associated with that subscriber's profile. It can then apply a different set of policies (e.g., throttling thresholds, permitted services) compared to the MNO's own direct subscribers, even though they are using the same physical radio infrastructure. This enables the host MNO to offer differentiated wholesale services. The SPI is also used in steering roaming scenarios to identify the home service provider of a roaming user.

Purpose & Motivation

The SPI exists to enable service differentiation and multi-tenancy in shared network infrastructures. The core problem it solves is the need for a single physical network (e.g., an MNO's RAN and core) to host multiple logical service providers (like MVNOs, enterprise customers, or IoT service providers) and treat their traffic differently according to commercial agreements. Without an SPI, the network could only apply policies based on the subscriber identity or APN, which is insufficient for complex wholesale and service aggregation models.

Historically, as mobile markets became more competitive, regulators and commercial pressures drove the rise of MVNOs. This created a technical requirement for the host MNO to identify which entity was ultimately responsible for a subscriber's service. The SPI, introduced in 3GPP Release 99, provided this capability. It was motivated by the need for flexible policy and charging control, which was becoming increasingly important with the advent of packet-switched data services (GPRS) where service tiers and billing models were more varied than simple voice calls.

It addresses the limitation of earlier networks that were essentially single-tenant. The SPI allows the decoupling of service provisioning from network operation. This enables new business models, such as an IoT platform provider (the service provider identified by the SPI) offering connectivity through multiple MNOs, with the MNOs applying the platform's specific policies uniformly. It is a cornerstone for enabling efficient and automated service delivery in modern, software-defined mobile networks.