Packet Loss Ratio

Description

Packet Loss Ratio (PLR) is a quantitative metric defined as (Number of Packets Lost) / (Number of Packets Sent), often expressed as a percentage or a value between 0 and 1. In 3GPP specifications, PLR is used both as a Key Performance Indicator (KPI) for network monitoring and as a parameter for defining Quality of Service (QoS) requirements. It is measured at various layers and interfaces: end-to-end for a service, over a specific bearer (e.g., an EPS bearer or QoS Flow), or across a particular network link (e.g., N3 interface in 5G). The loss can occur due to congestion, radio link errors, buffer overflows in network nodes, or handover failures.

From a technical measurement perspective, PLR is often calculated using sequence numbers in protocol headers. For example, in the RTP protocol used for voice and video, the receiver can analyze the sequence numbers of arriving packets to detect gaps and compute loss. In the 3GPP radio access network, lower layers like the Packet Data Convergence Protocol (PDCP) and Radio Link Control (RLC) have their own mechanisms to detect and report packet loss. The PDCP layer, responsible for header compression and integrity protection, maintains sequence numbering for user plane data. Counters for PDCP packet losses are defined for performance measurement (PM) in specifications like TS 38.314 (NR) and are reported to the management system. These measurements help isolate whether loss is occurring in the radio link versus the core network.

PLR is intrinsically linked to QoS management. In the 5G QoS model defined in TS 23.501, certain QoS Flow characteristics reference packet loss rate. For instance, a Guaranteed Flow Bit Rate (GFBR) QoS Flow for a Ultra-Reliable Low Latency Communication (URLLC) service will have a very stringent maximum allowed packet loss rate (e.g., 10^-5 or 10^-6). The network uses admission control and resource scheduling to try to meet this target. For multimedia services, 3GPP TS 26.114 defines performance objectives for IMS-based services, where PLR is a critical parameter for voice (VoLTE/VoNR) and video telephony. The network's transport layer, in conjunction with application-layer techniques like forward error correction (FEC) and retransmissions (e.g., RLC Acknowledged Mode for non-real-time data), works to keep the PLR within acceptable bounds for each service type.

Purpose & Motivation

PLR exists as a fundamental metric because packet loss directly and severely degrades the quality of experience (QoE) for digital services, especially real-time interactive ones. For voice over IP, even a small packet loss ratio (e.g., 1%) can cause audible clicks and gaps, while for video streaming, it leads to frozen frames or blocky artifacts. Quantifying this loss is the first step towards managing and mitigating it. Before standardized QoS metrics, network performance was assessed with vague terms; PLR provided an objective, measurable target for service level agreements (SLAs) and network optimization.

The motivation for its precise definition in 3GPP stems from the transition to all-IP networks in Release 5 and beyond, where traditional circuit-switched voice with its guaranteed delivery was replaced by packet-switched Voice over IP (VoIP). In packet networks, loss is inherent due to statistical multiplexing and best-effort routing. 3GPP needed to define acceptable loss thresholds for conversational services to ensure they remained viable over wireless IP networks. This led to the inclusion of PLR in QoS Class Identifier (QCI) characteristics in EPS and 5QI characteristics in 5G, allowing the network to treat traffic differently based on its loss sensitivity.

Historically, PLR addressed the limitation of not having a standardized way to correlate network performance with user-perceived quality. It enables several key network functions: 1) Performance Monitoring: Operators measure PLR to identify congested links or faulty equipment. 2) QoS Enforcement: The policy framework (PCRF/PCF) can use PLR measurements to trigger corrective actions. 3) Service Development: Application designers (e.g., for video conferencing) know the network's guaranteed PLR and can design appropriate error resilience (like FEC) around it. Thus, PLR is not just a measurement but a cornerstone for the entire QoS architecture in modern mobile networks.