Institute for Telecommunication Sciences / Research / 5G / IORS / 2025.06 IORS

IORS Virtual Meeting, June 2025

Key performance indicators and quality of experience

5G networks cover a great variety of use cases with different requirements, including high throughput (>1 Gbps), low latency (<1ms), and massive density of devices (106/km2). Some of these use cases involve intensive usage of video technologies, both from entertainment and from industrial perspectives. Each of those use cases may have different requirements in terms of network capacities or Key Performance Indicators (KPIs), such as throughput, latency, reliability, or density. Those factors will influence the final Quality of Experience (QoE) of the service, but not necessarily in a straightforward way.

This virtual meeting will present work underway in the Video Quality Experts Group (VQEG) 5G Key Performance Indicators group. Participants will then engage in a discussion. This meeting will expose differences between quality of service (QoS) metrics that directly measure the network performance and Quality of Experience (QoE) metrics that measure user experience. 

Logistics

  • Moderators:
    • Pablo Perez (Nokia)
    • Francois Blouin (Meta)
    • Kjell Brunnström (RISE Research Institutes of Sweden AB) 
  • Virtual meeting on June 11, 2025
  • Video recording available temporarily to registered attendees only 
  • Meeting slides

Agenda

  • Introduction to VQEG (Kjell)
    • Intro to 5GKPI project / VQEG Whitepaper (Pablo)
  • QoE definitions. From QoS to QoE (Pablo and Kjell)
  • Industry alignment. CAPs / CSPs challenges (Francois)
  • Towards a framework for QoE management (Pablo and Francois)

VQEG White Paper on Quality of Experience Aware Management for Collaboration between Network and Application Providers

This VQEG white paper (publication pending) addresses the challenge of improving end-user Quality of Experience (QoE) for Internet services. It begins by outlining the core problem: Content and Application Providers (CAP) and Communication Service Providers (CSP) operate largely independently without a common view of their users' experience. This separation makes it difficult to diagnose issues and optimize performance from an end-to-end perspective.

The white paper first establishes a common foundation by reviewing existing QoS and QoE definitions, QoE models and relevant industry standards. It presents a layered model to define key concepts, separating network-level Key Performance Indicators (KPIs), application-specific Key Quality Indicators (KQIs), and the user-centric QoE, proposing clear definitions for some important QoE-related terms, such as user-reported QoE, modeled QoE, or system QoE. This provides a common language for understanding the remainder of the paper and discussions in the research community.

The possible benefits of QoE management are discussed with respect to typical issues and common information gaps, making the case for closer collaboration between CAPs and CSPs. The core proposal is a framework for structured information exchange between those stakeholders. This mechanism, described as a shared state table, allows for the exchange of relevant metrics – either in near real-time or periodically, and with different granularity (e.g., in detail per user/media stream, or aggregated) – to create a shared view of service and network performance.

This exchange of information enables cooperative optimization. CSPs, on the one hand, can use QoE-related data from CAPs (e.g., video fidelity scores, stalling events) to better understand the impact of network conditions and adjust resource management accordingly. CAPs, on the other hand, can use network status information from CSPs (e.g., congestion levels, available throughput) to make network-aware adaptation decisions, such as selecting an appropriate video quality to avoid stalls.

To demonstrate the framework's real-world value, the white paper also illustrates its application through practical use cases for short and long-form video streaming, as well as interactive services like cloud gaming. These examples show how specific metrics can be used to improve startup times, reduce stalling, and manage latency. Finally, the framework addresses key privacy considerations, proposing a voluntary, opt-in system that uses practices such as temporary, pseudonymized session identifiers.

In conclusion, this VQEG white paper presents a structured approach for improving QoE through enhanced cooperation between CAPs and CSPs. The proposed framework provides a foundation for developing and sharing metrics that can lead to more efficient and effective service delivery. The recommended next steps include further development and validation of the proposed models through a proof-of-concept, with the long-term goal of contributing the findings to relevant standardization bodies such as ITU-T and IETF.