Communications Service Providers’ (CSP) mobile networks are major consumers of energy. A recent McKinsey report states that telecom operators already account for 2-3% of total global energy demand. The case for sustainable, aggressive improvements in energy efficiency and management is clear. #sponsored

August 18, 2022

6 Min Read
Why a RAN automation platform sets the best foundation for sustainability applications

Communications Service Providers’ (CSP) mobile networks are major consumers of energy. A recent McKinsey report 1 states that telecom operators already account for 2-3% of total global energy demand. The case for sustainable, aggressive improvements in energy efficiency and management is clear.

Largely due to the extensive geographical coverage they need to provide 2, A CSP’s Radio Access Network (RAN) typically consumes over 75% of the total energy used by the network. This large share means RAN efficiency is of the utmost importance in minimizing CSP’s environmental impact and keeping operating costs (OPEX) under control.

The foundation for improvements in RAN energy use lies in advances in improving both network hardware and software. In hardware, we see energy management as a focus in the ranges of systems on a chip (SoCs) that power the radios, as well as in novel ways to cool complex installations.

Software advances that reduce energy usage can be found throughout the network from core to radio. Of primary interest for this article will be evolutions in network automation, specifically orchestration and management for RAN.

RAN automation enabling sustainability success

There are great expectations in the market for a new approach to automation of the RAN domain as a number of key trends come together.

Communication networks have historically been driven by open standards, e.g. 3GPP, so the idea of aligning and supporting multi-vendor, multi-technology ecosystems is not new to the industry. As the central part of today’s technology defined world, this legacy of openness means telecommunications is the natural point of convergence for all parts of the wider technology industry. In effect, it can take the advances already seen in other areas and apply them where appropriate. Through openness, we take advantage of other trends like the improvement in data handling technologies which mean we can ingest and process large volumes of data at speed, and increasingly reliable and trusted Artificial Intelligence and Machine Learning (AI and ML) systems.

One of the initiatives regarding openness in telecommunications is Open RAN, with its focus on hardware/software disaggregation and interfaces based on open standards. Open RAN’s industry alignment includes the definition of the Service Management and Orchestration function (SMO) - the central point of configuration and orchestration for Open RAN resources. RAN automation applications (rApps) sit on top of the SMO, implementing specific automation use cases using the SMO as the gateway to the RAN.

A minority of network operations today use Open RAN. So, acknowledging the benefits of the mentioned technology trends, it makes sense to apply these to the wider RAN, not only in Open RAN deployments. This can be done through open interoperable interfaces. This is why we think advances in energy efficiency will require multi-vendor and multi-technology support from the ground up. This approach would help deliver the capabilities expected to be embedded in SMO through RAN automation applications that operate across Open RAN as well as existing purpose-built networks, resulting in better energy management for reasons we may summarize as: “more data”, “faster data”, “enriched insights” and “central AI/ML capabilities”.

rApps and energy management

The rApps, for either SMO and other automation platforms with similar capabilities, should be highly functional applications, benefitting from open platform capabilities and advances in data management and analysis, to deliver the energy management use cases.

A typical rApp would retrieve data from the orchestration platform to understand the status of the Radio network, both in topology and performance aspects, and use AI / ML models to decide which is the best network configuration possible to reach the objective. In the case of energy management rApps, that would be usually reducing power consumption without affecting customer experience or performance.

Proven impact of sustainability applications

While rApps will soon embody the automation use cases in the SMO architecture, many of the use cases are already implemented and being delivered in other platforms, although not yet benefiting from the increased value and performance the SMO can offer. Among the use cases already providing value to operators globally we can mention Power Optimization and Predictive Cell Energy Management applications.

Power Optimization uses digital twins to model and recommend the optimal radiated power settings for a cell (Figure 1), ensuring regulations on radiated power are met while protecting coverage and user experience. Recent customer trials with Swisscom 3 delivered an average downlink power reduction of 20% which translates into 1.7% reduction on overall site level energy usage.

Figure 1: Figure 1. Ericsson Power Optimization application with reinforced learning Figure 1. Ericsson Power Optimization application with reinforced learning

Another example, Predictive Cell Energy Management, uses an ML prediction model operating across different vendors and technologies, accounting for traffic, interference, and optimal user experience, to decide when to stop and restart each cell’s transmission across the entire network. In our live deployments and trials, we are able to reduce network consumption between 2% to 8% in addition to the effect of energy efficiency RAN software features, accounting for millions of dollars saved yearly in the energy bill and thousands of tons of CO2 emissions prevented. In the recent case of an Asian Tier-1 with over 350 million subscribers, the savings potential amounts to 15 MUSD/year and 100 KTons of CO2 emissions across the full network.

An rApp, operating across SMO and other automation platforms with similar capabilities, will deliver benefits such as:

  • Flexibility and innovation with the open ecosystem allowing us to bring development from other industries into the telecom industry

  • Enhanced functionality with the use of multiple data sources and AI/ML techniques

  • Consolidation in one automation platform that allows us to orchestrate and coordinate actions towards the network

Looking towards a sustainable future

The ongoing energy challenges will accelerate adoption of sustainability focused applications. We see new platforms based on AI and ML, like the SMO and rApps, as fundamental enablers of the evolution of these important applications.

Ericsson’s ambition is to provide high quality, high performing 5G experiences while simultaneously aiming to help operators around the world reduce network energy consumption. To this end, Ericsson has developed a network-wide approach called “Breaking the energy curve”. It provides a holistic approach to introducing 5G while managing mobile network energy use across core, transport, radio access and site equipment. Automation, such as we see in rApps, is a crucial element in tying together multiple streams of technology development to ensure both Ericsson and its customers worldwide are successful in creating environmentally friendly, high-performance communications networks.

1 McKinsey: The case for committing to greener telecommunications networks , Feb 2020
2 Ericsson: Sustainability & Corporate Responsibility Report - Ericsson
3 Ericsson: Applying AI in telecoms – Mobility Report - Ericsson

This content is sponsored by Ericsson.

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like