Energy efficiency has become an increasingly important consideration in mobile network design and operation. As networks support higher traffic volumes, more demanding applications and growing AI workloads, operators need to balance network performance, user experience and energy consumption.
We have previously looked at this topic in our post on 3GPP Release 19 Description and Summary of Work Items, where energy efficiency and energy saving featured prominently. More recently, we produced a five-part video series on Energy Efficiency Improvements in 3GPP Technologies, tracing the evolution of energy-saving mechanisms from LTE through 5G NR and 5G-Advanced Releases 18 and 19.
Much of the earlier work concentrated on improving the energy efficiency of the network itself. This includes mechanisms such as cell discontinuous transmission and reception (DTX/DRX), antenna adaptation, on-demand signalling, low-power wake-up receivers and other techniques that allow network equipment and devices to reduce energy consumption without unnecessarily affecting service performance.
An interesting development in Release 20 is the increasing attention being given to energy efficiency at the application layer. Rather than focusing exclusively on how the network consumes energy, 3GPP is examining how applications can use energy-related information to make more efficient decisions about where, when and how services are delivered.
This work is described in 3GPP TR 23.700-44, Study on Application Enablement to Support Energy Saving Phase 2. The July 2026 version, V20.1.0, documents the work undertaken by 3GPP SA6 to investigate application-layer energy-saving capabilities and identifies solutions for subsequent specification work.
One of the key concepts is Energy Saving Enablement (ESE), which introduces common capabilities for collecting, processing, monitoring and exposing energy-related information to applications. The proposed architecture includes ESE servers and clients, allowing applications and other service enablers to obtain energy information and receive notifications when relevant conditions change.
This builds on existing capabilities in the 5G Core and network management systems. The objective is to make energy-related information more useful to applications rather than leaving it exclusively within network management and optimisation functions.
One interesting application is energy-aware edge computing. Traditionally, edge application servers are selected primarily using considerations such as location, latency, service availability and performance. The Release-20 study explores how energy information, including renewable energy availability and energy-related operating characteristics, could also influence server selection.
For example, where multiple edge servers can provide the required service, an application could potentially select one based on its energy characteristics, provided that performance and other requirements continue to be satisfied. The study also considers how energy-related information could influence Quality of Service (QoS) decisions.
Another important area is Artificial Intelligence and Machine Learning (AI/ML). As AI processing becomes more distributed across networks, edge infrastructure and devices, the energy consumed by training and inference becomes increasingly relevant.
The study examines enhancements to AI/ML enablement services that could support energy-aware client selection, task transfer and split AI/ML operations. In federated learning, where multiple devices participate in training a shared model without transferring all their local data to a central location, energy information could help determine which devices participate in a particular training activity.
It also investigates analytics capabilities for monitoring or estimating the energy consumption associated with AI/ML operations. This is particularly interesting as networks increasingly support distributed intelligence and AI-based applications.
Network slicing is another area covered by the study. Existing network slicing mechanisms allow different services and applications to receive network resources appropriate to their requirements. The Release-20 work considers how energy-related information could influence slice selection, modification and adaptation, including coordinated optimisation across multiple slices.
However, energy saving cannot be considered in isolation. Any such optimisation must take account of service requirements, including latency, reliability and performance. This is particularly important for industrial, enterprise and mission-critical applications where reducing energy consumption must not compromise operational requirements.
The report also examines energy-aware location services. Location reporting can consume energy, particularly when frequent updates are required. The proposed enhancements consider how location reporting configurations could be adapted using energy-related analytics and operating conditions, potentially reducing unnecessary reporting while continuing to meet application needs.
An important distinction throughout this work is that energy efficiency, energy consumption and sustainability are related but different objectives. Selecting an edge server using renewable electricity, for example, may reduce the carbon footprint associated with a service without necessarily reducing the total energy consumed. Similarly, reducing the frequency of an operation may save energy but could also affect service performance.
The Release-20 work therefore represents an effort to make energy information available across more parts of the system, enabling better-informed decisions rather than assuming that every energy-saving mechanism is suitable for every application.
Importantly, this work has progressed beyond the initial study. A new specification, 3GPP TS 23.366, Functional Architecture and Information Flows for Energy Saving Enablement Service, reached version 20.0.0 in September 2026. It provides the normative architecture and procedures for the Energy Saving Enablement service. Related application-layer enhancements are addressed in other Release-20 specifications.
It is worth remembering that the technical report contains several candidate solutions and recommendations, not all of which necessarily appear unchanged in the final specifications. Standardisation also does not mean these capabilities are already available in commercial networks.
Looking at the broader evolution, LTE introduced practical network energy-saving mechanisms, 5G NR provided a more flexible and energy-efficient foundation, and 5G-Advanced Releases 18 and 19 expanded energy optimisation across radio, device and system functions. Release 20 is extending this work towards application-layer energy awareness.
As AI workloads, edge computing and increasingly diverse applications become more common, the ability to consider energy consumption alongside conventional performance requirements could become an important part of future network and service optimisation.
Related Posts:
- The 3G4G Blog - Tutorial: A Quick Introduction to 3GPP
- The 3G4G Blog: 3GPP Release 19 Description and Summary of Work Items
- The 3G4G Blog - Energy Efficiency Improvements in 3GPP Technologies: From LTE to 5G-Advanced
No comments:
Post a Comment