Data IDentification in ANDSF

Description

DIDA (Data IDentification in ANDSF) is a functional component integrated into the Access Network Discovery and Selection Function (ANDSF) framework defined in 3GPP Release 11 and subsequent releases. Its primary role is to provide a granular identification mechanism for data traffic flows or specific applications traversing a User Equipment (UE). DIDA works by defining data identification rules and containers within ANDSF management objects (MOs). These rules can be based on various packet flow descriptors such as destination IP addresses, port numbers, protocol types, or application identifiers (e.g., from deep packet inspection or OS-provided signals). The ANDSF server, typically operated by the network provider, can provision these DIDA policies to the UE over an S14 interface.

When a UE receives DIDA policies, its ANDSF client stores and interprets them. As data packets are generated by applications, the UE's policy enforcement point matches these packets against the active DIDA rules. A successful match associates the data flow with a specific DIDA identifier. This identifier is not used for routing itself but serves as a key. The core function of DIDA is to act as a selector for other ANDSF policies, namely ISRP (Inter-System Routing Policy) and IARP (Inter-APN Routing Policy). These routing policies can be configured to apply only to data flows that have been tagged with a particular DIDA identifier.

Architecturally, DIDA introduces new nodes within the ANDSF management object tree, specifically the `<DIDARule>` and `<DIDAContainer>` elements. A `<DIDARule>` contains the actual matching criteria (like a traffic flow template), while a `<DIDAContainer>` groups multiple rules and assigns a unique DIDA ID. This hierarchical structure allows for complex, multi-criteria identification. The process is integral to policy-based traffic steering, enabling scenarios where, for example, all video streaming traffic (identified via DIDA) is routed to a Wi-Fi network while keeping latency-sensitive gaming traffic on the cellular link, even if both originate from the same IP address block.

DIDA's role in the network is that of an enabler for intelligent access selection and traffic management. By moving beyond network-level or APN-level steering, it allows operators to implement service-aware policies. This granularity helps optimize network resource utilization, manage congestion, and implement service-specific charging or quality-of-service treatments. It forms a foundational piece for later enhancements in multi-access connectivity and seamless offload, providing the necessary 'data awareness' for sophisticated ANDSF-driven decisions.

Purpose & Motivation

DIDA was introduced to address the growing need for granular, application-aware traffic steering in heterogeneous networks (HetNets). Prior to Release 11, ANDSF policies primarily operated at a broad level, steering all traffic from a UE or all traffic associated with a specific Access Point Name (APN) towards a particular access network (e.g., Wi-Fi or 3GPP). This coarse granularity was insufficient as smartphones proliferated and user traffic became a mix of diverse applications with vastly different requirements (e.g., video streaming, VoIP, web browsing, software updates). Operators lacked the tool to differentiate between these flows and apply distinct routing policies.

The creation of DIDA was motivated by the desire to implement service-based network selection and offload. Without DIDA, an operator wanting to offload only video traffic to a complementary Wi-Fi network while keeping voice calls on the cellular network had no standardized way to identify the 'video traffic' flow within ANDSF. DIDA solved this by providing a standardized container for flow identification criteria that could be dynamically provisioned. It addressed the limitations of previous, non-standardized or proprietary deep packet inspection (DPI) solutions embedded in the network, which were complex, costly, and often raised privacy concerns. DIDA shifted the intelligence and the matching logic to the UE side, based on operator-provided rules, creating a more scalable and privacy-friendly model.

Historically, DIDA's development aligned with the industry's move towards intelligent traffic management and the early concepts of network slicing and service-aware architectures. It provided the foundational 'filtering' mechanism that allowed ANDSF to evolve from a simple access discovery tool into a sophisticated policy engine for Multi-Access PDN Connectivity (MAPCON) and later, IP Flow Mobility (IFOM). By solving the data identification problem, DIDA unlocked the potential for truly dynamic, policy-driven multi-access networks, paving the way for the seamless integration of non-3GPP accesses envisioned in 5G.

Key Features

• Defines standardized rules for identifying specific data flows or applications at the UE.
• Uses traffic flow templates (TFT-like) with matching criteria such as IP address, port, and protocol.
• Provides a hierarchical structure with DIDARules and DIDAContainers for organized policy management.
• Serves as a selector key for Inter-System Routing Policy (ISRP) and Inter-APN Routing Policy (IARP).
• Enables application-aware or service-specific traffic steering across 3GPP and non-3GPP accesses.
• Supports dynamic provisioning and update from the ANDSF server to the UE via the S14 reference point.

Evolution Across Releases

Defining Specifications