As described in the use cases summarised in clause 5.1, energy-related information needs to be collected, estimated, and exposed by the 5G System. This information is not only necessary for internal network optimisation, but it will also benefit service adaptation by third-party Application Service Providers. With the consent of Mobile Network Operators, it is relevant to expose energy-related information (e.g., energy consumption information, energy efficiency information, renewable energy and carbon emission information) to authorised consumers. TR 23.700-66 study and identify potential enhancements to the 5G System (e.g. including network energy related information exposure, enhancement for subscription and policy control to enable energy efficiency as service criteria) to improve energy efficiency and to support energy saving in the network, taking the Energy requirements in TS 22.261 into consideration. The purpose of this Key Issue is to extend this work, to study and identify potential enhancements to the energy-related information exposure for media architectures, applications and services (e.g., taking into consideration the 5G Media Streaming System according to TS 26.501, 5G Multicast-Broadcast User Services according to TS 26.502, the Real-time Media Communication according to TS 26.506, Split rendering Media session Enabler according to TS 26.565, etc.) including UE-related energy information exposure. As explained in clause 4.2.2, a UE data collection and reporting architecture already exist in the 5G System. But energy-related information is missing. In this context, the subsequent analysis by this Key Issue should consider:

1. How should UE energy-related information be reported by a UE to the 5G System?
2. Which reference points should be used to report UE energy-related information to the Data Collection AF?
3. Would it be useful to expose energy-related information of the network to the Media Session Handler to help it optimize its media session in an energy-efficient way?
4. How to allow a UE to report its energy-related information without exposing its energy consumption rate?

This analysis will need to take into consideration existing work done in 3GPP but also other market trends. Application Service Providers are often reluctant to deploy solutions specific to mobile networks. Proposing technologies already supported in their services or technologies that are agnostic to the network, or which operate passively without the active involvement or knowledge of applications, are more likely adopted.

Clause 6.4 in TR 22.882 contains the consolidated requirements extracted from use cases, related to information exposure related with this Key Issue: [CPR 6.4-1] [CPR 6.4-2] [CPR 6.4-3] [CPR 6.4-4] [CPR 6.4-5] Additional potential requirements identified from related work: [PR 1-1] [PR 1-2] [PR 1-3]

The measurement of energy consumption by media services is an essential pre-requisite to facilitate efficient energy use and energy saving. The environmental impact of the consumption of media services depends various system actors: equipment manufacturers, networks, datacentre operators, service providers but also users through their usage behaviour. Requirements of monitoring and measurement relating to energy consumption and efficiency have already been defined by 3GPP in TR 22.882 but those requirements are only related to energy consumption monitoring of the 5G network or 5G System. TR 23.700-66 studies and identifies potential enhancements to the 5G System (e.g. including network energy related information exposure, and enhancement of subscription and policy control to enable energy efficiency as a service criterion) with the dual aims of (i) improving energy efficiency and (ii) supporting energy saving in the network, taking the EnergyServ requirements in TS 22.261 into consideration. The purpose of this Key Issue is to extend the above work, by studying and identifying potential enhancements to energy-related monitoring and measurement aspects (including UE-related energy information measurement and monitoring) in the context of the following media delivery architectures, applications and services namely the 5G Media Streaming System according to TS 26.501, 5G Multicast-Broadcast User Services according to TS 26.502, the Real-time Media Communication according to TS 26.506, Split rendering Media session Enabler according to TS 26.565, etc. The most efficient way to address energy saving in this context is to consider the complete end-to-end media delivery chain, seeking to understand how decisions made on one end of the chain impact power consumption on other ends of the workflow. As explained in clause 4.2.2, UE data collection has been defined to monitor QoE in the 5G system. But energy-related information is missing. In this context, the subsequent analysis by this Key Issue will consider the following questions:

1. Which UE energy-related information will be collected to measure, correlate, and optimize energy usage across the entire streaming distribution chain?
2. Can existing methods be leveraged to measure/monitor the identified UE energy-related information?
3. Which UE entity is appropriate to measure this UE energy-related information?

This issue is even more important for advanced media services such as XR services, Split Rendered media services, etc. which are expected to incur substantial energy consumption both at the device and network levels, presenting significant challenges for operators and service providers. For instance, when a Mobile Network Operator (MNO) deploys a communication service to fulfil application service requirements, such as those of a gaming application, it is crucial for the customer - whether an Application Service Provider (ASP) or an industry vertical - to ensure that the application service reduces energy consumption for both end users and the data network. Failure to do so could necessitate short-notice application layer adaptations within the Application Service Provider's domain. This may result in adjustments being made to service levels in response to anticipated high energy consumption in specific service areas or during peak hours that adversely affect the Quality of Experience for service users. In a practical scenario, an ASP intends to deploy a gaming service within a designated service area, served by an MNO's 5G network. Various service levels may exist, each associated with specific Key Performance Indicators (KPIs), such as automation levels or video quality targets. To ensure energy efficiency the ASP monitors the energy efficiency of its application service across specified service levels. Monitoring of application energy consumption may occur periodically or may be event-triggered, depending on the ASP's requirements, which are typically outlined in the Service Level Agreement (SLA). The MNO and ASP may agree on a certain energy efficiency target for the application service and optionally for given service levels. This analysis will need to take into consideration existing work done in 3GPP but also other market trends. Application Service Providers are often reluctant to deploy solutions specific to mobile networks. Proposing technologies already supported in their services or technologies that are agnostic to the network, or which operate passively without the active involvement or knowledge of applications, are more likely adopted.

Clause 6.3 in TR 22.882 contains the consolidated requirements extracted from use cases, related to monitoring and measurement related with this Key Issue: [CPR 6.3-1] [CPR 6.3-2] [CPR 6.3-3] Based on TS 28.310 and other related work in 3GPP, the following potential requirements need to be considered in this Key Issue: [PR 2-1] [PR 2-2] [PR 2-3]

The complexity of the media delivery chain raises technical challenges in terms of environmental impact measurement and accountability. There is no explicitly and uniformly defined methodology to evaluate the energy usage/savings of multimedia standards features and proposals which leads to significant variation in the results of different studies. Regulators and legislatures like the European Parliament have already provided more precise guidelines, targeting companies or services, to assess cases of unfair environmental advertising [60]. A characterisation framework to evaluate the energy usage/savings of multimedia standards features and proposals for relevant 3GPP scenarios is also needed to justify claims. The framework presented in this Key Issue is based on the normative work (e.g., granularities or energy-related information) recommended by Key Issues #1 and #2 of the present document, and includes a set of relevant scenarios and anchors. It is not dependent on a particular implementation of the feature or other parts of the system and allows an energy-efficient implementation to be compared against an implementation not relying on the proposed standards feature. The commonly supported external estimation frameworks mentioned in clauses 4.2.3 and 4.2.4 are to be reviewed and agreed by SA WG4 before using them to develop the evaluation framework in this Key Issue. Characterization of specific media-related technologies has previously been conducted by 3GPP, for example in TR 26.955 on Video Characterization and TR 26.947 on Characterization of Application Layer FEC. Drawing from these frameworks, a foundation for evaluating the energy usage and savings of specific multimedia technology features may be derived. Some best practices from these earlier characterization frameworks can be summarized as follows:

1. Define representative and relevant scenarios: The evaluation of technologies is expected to be done by selecting scenarios that are typical and representative for the usage of a technology in the context of 3GPP services or media delivery context.
2. Establish clear anchors and metrics: Anchors are used to provide a reference technology, for example an already defined technology in 3GPP (for example H.265/HEVC in the context of TR 26.955 or Raptor codes in the context of TR 26.947). Metrics need to provide a meaningful indication of the impact of the technology for the desired objective. Preferably, these metrics are derived by well-defined and deterministic measurement.
3. Ensure repeatability: In order to ensure credibility of the anchor and test results, the evaluation is preferably repeatable by an independent evaluator. Repeatability may, for example, be ensured by well-defined reference software.
4. Make technology tests meaningful: If new technologies are evaluated against anchors, the executed tests are expected to be meaningful and comparable to the anchors. Tests also are expected to provide relevant information in the context of the scenario.
5. Provide a reference implementation: In order to provide consistency and repeatability, a reference software implementation that is openly accessible is a beneficial asset.
6. Define a verification and validation process: In order that results have the highest credibility, providing an independent validation framework of them is encouraged, as well as documenting independently verified results. If the results cannot be verified, the validation framework provides the means to identify bugs or inconsistencies and potentially fix them.
7. Define an example characterization: To ensure that the framework is consistent, providing an example characterization is encouraged, i.e. the comparison of defined anchor results with some well-known comparable technology.
8. Work is consensus-based and reason-based: Comparable configurations between anchors and tests are not necessarily easily defined. It is expected that experts can reach consensus on reasonable and comparable configurations. Also, determined metrics may provide objective results, but the interpretation of the results is expected to be based on reasoned arguments, and results should always be viewed in the context of how they were determined.
9. Avoid mixing solution selection and characterization - focus on the latter: Expanding on point 8 above, characterization results primarily support the discussion to evaluate a technology, but are typically not suitable for a technology selection shoot-out because such an ambition may limit consensus-based discussions.
10. Do not strive for a single KPI number; instead collect information to have a basis for argumentation: Also, expanding on point 8 above, the collection of multiple data points allows experts to use the numbers in argumentation. Therefore, more independent metrics support consensus building.
11. Initially, prioritize only one or two scenarios for a more focused approach: Scenarios are typically complex to establish, and hence too many scenarios may not be manageable. A reasonable set of scenarios that are well executed is preferable to an overload of scenarios.In addition to the above best practices, the candidate technology under evaluation will first have to overcome complexities already encountered with past characterization work:
    • Getting good and meaningful and agreeable scenarios and anchors.
    • Implementation dependency.
    • Documentability.
    • Verification and validation: cross validation, test labs, etc.

[PR 3-1] [PR 3-2]