Performance Measurement

Description

Performance Measurement (PM) is a cornerstone of the 3GPP Telecommunications Management Network (TMN) and Network Management (NM) framework, primarily specified in the 32-series (Management and orchestration) and 28-series (Management and orchestration for 5G networks) technical specifications. It defines a comprehensive system for gathering raw performance data, known as Performance Measurements, from managed network elements (NEs) like NodeBs, eNodeBs, gNBs, and core network functions. These raw measurements are counters, gauges, and status indications that reflect the operational state and load of the network resources.

The architecture involves Performance Measurement Jobs (PM-Jobs) configured by the Management System (e.g., Operation Support System - OSS) on the Network Elements or Element Managers. A PM-Job specifies what measurements to collect, the granularity (e.g., 15-minute, 1-hour, 24-hour), and the schedule. The Network Element, acting as a measurement producer, collects the raw data over the specified granularity period, resulting in a set of PM data. This data is then formatted into standardized PM files (e.g., XML-based) and transferred to the management system via file-based interfaces like FTAM or more modern mechanisms.

Upon receipt, the management system, acting as a measurement consumer, processes these PM files. Processing includes parsing, validation, aggregation (e.g., rolling up cell-level measurements to site-level), threshold crossing analysis, and correlation with other management data like Fault Management (FM) alarms. The processed data is stored in a Performance Management (PM) repository for historical analysis and reporting. Key performance indicators (KPIs), which are derived from one or more raw PM counters using defined formulas, are calculated to provide actionable insights into network performance, such as call setup success rate, handover success rate, throughput, and latency.

The role of PM is critical for the entire network lifecycle. It provides the empirical data needed for capacity planning, identifying congestion hotspots, troubleshooting service degradation, verifying Service Level Agreements (SLAs), and driving automated network optimization and Self-Organizing Network (SON) functions. In 5G and beyond, PM data also feeds into network analytics, AI/ML applications, and closed-loop automation for intent-based management.

Purpose & Motivation

Performance Measurement was created to address the fundamental operational need for quantitative, objective data on network behavior. Prior to standardization, vendors implemented proprietary performance monitoring solutions, making it extremely difficult for operators with multi-vendor networks to get a unified, consistent view of end-to-end performance. This fragmentation hindered efficient network operations, troubleshooting, and service quality management.

The primary problem PM solves is providing a vendor-neutral, technology-agnostic framework for collecting performance data. This allows operators to deploy equipment from different manufacturers while maintaining a single, coherent operational view. It enables benchmarking, consistent KPI definition across the network, and fair assessment of equipment and service performance. The historical context lies in the broader TMN principles, where FCAPS (Fault, Configuration, Accounting, Performance, Security) management is a key model. PM fulfills the 'Performance' pillar, providing the data necessary for proactive and reactive management.

Furthermore, as networks evolved from voice-centric 2G to data-rich 5G and IMS, the complexity and types of performance metrics expanded dramatically. The standardized PM framework ensured that new metrics for packet data throughput, IP Multimedia Subsystem (IMS) session quality, and 5G network slicing performance could be seamlessly integrated into the existing management infrastructure. It provides the foundational data layer for evolving operational paradigms like zero-touch network and service management.