YOU ARE AT:AI-Machine-LearningAI is key to building high-performing, energy-efficient mobile networks (Reader Forum)

AI is key to building high-performing, energy-efficient mobile networks (Reader Forum)

Driven by an ever-expanding universe of connected devices and more powerful smartphones, mobile data traffic is continuing to surge, meaning the stakes are higher than ever when it comes to providing high-performing 5G networks.

For years, the “best network” was defined as the one that was the fastest and most reliable. But as communication service providers (CSPs) race to meet science-based climate reduction targets, including Net Zero goals, and focus on securing network energy performance, expectations have shifted — performance is no longer the only standard. And it’s not just the carriers themselves who have this focus — a survey by GSMA Intelligence found that 30 to 60% of mobile customers would be willing to pay more for a product certified as carbon neutral. It’s time to redefine the “best network” as not just high-performing, but energy efficient and sustainable as well.

For network operators, this presents unique challenges. In the coming years, mobile data traffic across the globe is only expected to grow. In fact, Ericsson’s recent mobility report indicates traffic increasing by four times from 2022 to 2028 in the US alone. As a result, efficient networks are more critical than ever.

As we work to scale 5G and support applications like automated vehicles, smart cities and connected healthcare, 5G technology promises groundbreaking use cases and better service for customers. CSPs have to strike the right balance between scaling 5G to meet consumers and business expectations, while managing increasing mobile data traffic and aiming to secure network energy performance and achieve Net Zero emissions. It’s a tall task, that can only be achieved with a holistic approach to network energy management.

Achieving energy efficiency with 5G

By design, 5G is the most energy aware standard yet. But to truly support the revolutionary use cases like augmented reality that depend on 5G’s high speed and low latency, more net new 5G sites are being deployed. The challenge for communications services providers becomes how to offer better performance and more capacity with less energy consumption while using lighter infrastructure. Operators can’t afford to be seen as compromising quality in pursuit of saving energy costs.

Without an energy management strategy at every level, network power consumption will grow, increasing costs and making it more difficult to achieve carbon emission reduction goals. Only by considering every aspect of the network, from precision deployment to modernizing hardware to AI-driven software solutions and intelligent site infrastructure, can CSPs truly achieve energy efficiency while meeting rising user expectations for quality of experience.

Hardware built for high traffic

The first step to creating a more sustainable network is to focus on hardware, as outdated equipment is simply not efficient. In fact, the latest 5G equipment is capable of achieving ten times improved capacity and more than 30% energy savings.

In high traffic scenarios, Ericsson has found that delivering 5G data with Massive MIMO radios is 13 times more energy efficient than LTE FDD (Frequency Duplex Division). Modernizing hardware equipment alone brings significant energy savings — but to truly unlock energy efficiency, operators need to focus on 5G’s energy-saving software features, too.

AI-enabled tools are driving sustainable 5G

As 5G networks are deployed, they require closed loop orchestration using real time data to adjust to patterns of high usage and demand. As cell networks become denser and more complex, there is really only one way to do this at scale — with AI.

There are solutions currently in the marketplace that help networks become more energy efficient. But without AI enabled tools, the gains are limited to things like turning equipment off at specific times e.g. at nighttime or during traffic variations.

Think of it like a programable thermostat — setting two separate temperatures for “home” and “away” can save a homeowner a few dollars in energy bills. But an AI-powered, constantly learning thermostat can analyze patterns and identify substantial opportunities to reduce energy consumption without compromising human comfort.

In the same way, AI tools in a network can bring major energy efficiencies to a cellular network, beyond optimizing for peak traffic, with minimal impact on quality of experience for users. It’s what will take energy orchestration from reactive to predictive. To further refine this, “intent-based networks” will allow service providers to control the automated functions in a closed loop by defining the most important network behaviors in advance, which the AI-tools will then translate into action across the system with minimal human intervention. To return to the thermostat metaphor, it will be like a smart thermostat optimizing for the most comfortable temperature while at the same time, looking for energy savings.

Efficient monitoring and tracking supported by AI has been shown to reduce RAN energy use by up to 12% annually. Because AI programs dynamically learn, adapt and act accordingly, these tools enable operators to control cells dynamically and in turn, serve dynamic traffic patterns instead of just peak traffic, making them a crucial component in boosting efficiency without sacrificing performance or risking service outages.

Redefining the “best network”

As CSPs look to meet Net Zero goals and manage energy usage for dense radio deployment in 5G networks — while meeting continually high expectations for quality of experience — it will take every tool in the toolkit to find success. An energy efficient standard in 5G, modernized efficient hardware and AI tools that leverage predicative features to discover energy savings will all be critically important. Moving forward, energy savings will become an important measuring stick along with network performance, as CSPs work to bring the best network possible to their customers.

ABOUT AUTHOR