Packet Set Error Rate

Description

Packet Set Error Rate (PSER) is a statistical QoS metric defined within the 5G Quality of Service (QoS) framework. It quantifies the reliability of packet delivery for applications where data is organized into logical groups or 'packet sets'. A packet set is a collection of packets that are semantically related and have a collective delivery deadline. The PSER is calculated as the ratio of packet sets that experience at least one packet delivery failure (e.g., loss, excessive delay beyond a latency bound, or out-of-order delivery that violates application constraints) to the total number of packet sets transmitted.

The metric operates in conjunction with other QoS parameters like Packet Delay Budget (PDB) and Packet Error Rate (PER). However, PSER provides a more holistic view for applications where the integrity of an entire data unit (the set) is paramount. The network uses the PSER target, along with the associated latency bound (often a maximum allowed packet delay), to make resource allocation and scheduling decisions. For instance, for an Augmented Reality stream, a frame might be composed of multiple packets; the PSER ensures the entire frame is delivered correctly and on time, as the loss of even one packet could degrade the user experience.

Architecturally, PSER is part of the 5G QoS Indicator (5QI) table and associated QoS Flow descriptions. It is signaled between the User Equipment (UE), the Radio Access Network (RAN), and the Core Network (specifically the Session Management Function - SMF) to establish the required service characteristics. The RAN scheduler is a key component responsible for meeting the PSER target, employing techniques like prioritized scheduling, redundancy (e.g., through packet duplication), or hybrid automatic repeat request (HARQ) strategies tailored for set-based reliability.

PSER's role is pivotal for enabling deterministic communication in 5G-Advanced and beyond. It moves beyond per-packet guarantees to per-data-unit guarantees, which is essential for time-sensitive applications. Its specification across documents like TS 23.501 (system architecture), TS 38.300 (RAN overall), and TS 38.835 (work on NR industrial IoT) highlights its integration into both the core service definition and the radio interface procedures.

Purpose & Motivation

PSER was created to address the stringent and complex reliability requirements of emerging 5G-Advanced and 6G-era applications, particularly extended reality (XR), cloud gaming, and ultra-reliable low-latency communication (URLLC) for industrial automation. Traditional QoS metrics like Packet Error Rate (PER) and Packet Delay Budget (PDB) provide guarantees on individual packets but do not adequately capture the performance needs of applications where data is structured in groups (e.g., a video frame, a control command with multiple parameters, or a synchronized sensor burst). A single packet loss within such a group could render the entire data unit useless, even if the average PER is low.

The motivation stems from the limitations of previous approaches. Guaranteeing a low PER for every packet is often inefficient and resource-intensive. PSER allows the network to optimize resource usage by focusing on the successful delivery of logically complete sets within a time window, providing a more application-aware reliability metric. This enables network slicing and QoS frameworks to support services with 'bounded reliability' – a concept where occasional set failures might be tolerable if bounded within a statistical limit, allowing for more flexible and efficient network operation compared to traditional 'five-nines' (99.999%) per-packet reliability models that are often over-provisioned.

Historically, as 3GPP evolved from enhanced Mobile Broadband (eMBB) to massive IoT and critical IoT, the need for more granular and diverse QoS metrics became apparent. PSER, introduced in Release 18 as part of the work on enhanced XR and industrial IoT, represents a significant step in refining the QoS toolbox. It allows service providers to define and guarantee Service Level Agreements (SLAs) that directly map to the perceptual or operational quality of the end-user application, facilitating new revenue streams and enabling truly deterministic wireless performance for industry 4.0.

Key Features

• Set-based reliability metric measuring failure rate of logical packet groups
• Defined in conjunction with a maximum latency bound for the packet set
• Integrated into the 5G QoS Framework and 5QI table
• Enables application-aware resource scheduling in the RAN
• Supports statistical bounded reliability for efficiency
• Critical for deterministic performance in XR and industrial IoT applications

Evolution Across Releases

Defining Specifications