Wireless networks have always consumed a lot of power. However, compared to earlier mobile technologies, 5G offers a far higher number of simultaneous connections and a higher data throughput. Although 5G has been designed with greater intrinsic network efficiency than 4G, it uses more energy because of its capacity to sustain significantly higher traffic loads and handle more complex tasks more quickly.
According to research from GSMA Intelligence, for instance, the average customer data usage on 5G networks is generally five to ten times higher than that of LTE. Researchers anticipate that by 2025, there will be over 7.5 billion mobile subscribers worldwide. Although the 3GPP 5G standards provide certain automated power management features, 5G networks will result in considerable increases in power usage unless MNOs take thorough measures to manage it.
Managing power while maintaining SLAs
Automation management solutions that minimize output while maintaining service-level agreements (SLAs) are frequently needed to control power consumption and its environmental impact. In 5G enterprise applications, which are frequently mission-critical and depend on real-time connectivity and automation, upholding service levels is crucial.
It’s possible to control the complexity and density of 5G traffic by using AI-driven, predictive analytics to anticipate network traffic patterns and automatically make adjustments to achieve optimal power-performance balance. Smart, self-regulating networks made possible by AI are able to detect anomalies in network activity and adjust power levels to reduce output while preserving performance. Since traffic patterns will vary greatly from day to day and hour to hour as 5G applications become more dynamic, this function goes beyond applying fixed, status shutdown windows based on previous actions or network activity.
According to some estimates, an AI-based energy management system can cut the energy consumption and cost of MNOs’ networks by 20% to 30% without degrading performance. These cuts represent two to five times higher savings than non-AI systems that perform temporary shutdowns based on predefined schedules. In highly distributed 5G infrastructures, predictive systems can provide small-scale power savings that, over time, could build up to more significant ones.
RAN with the lion’s share of power
The radio access network (RAN) accounts for an estimated 80% of mobile network consumption. Analysts also predict that 5G base stations will consume at least double the amount of power as 4G base stations. Therefore, the RAN is the most sensible starting point for MNOs in terms of enhancing energy efficiency.
Within each RAN, the flow of traffic varies during the day. The most efficient energy management makes use of AI to dynamically and automatically turn off specific RAN components when not in use in order to save energy and reduce costs.
Artificial intelligence can give priority to critical 5G users and/or network slices in order to achieve the best power-SLA balance. Operators may provision thousands of virtual, private 5G enterprise networks on their existing 5G infrastructures thanks to 5G network slicing. Each slice, which represents a particular customer network, application, or location, can have its own SLA. To maintain adherence to the crucial SLA, the AI-driven energy management system can activate a dormant base station or boost power levels in one that has been scaled back in the event of network congestion or a failure. This is comparable to adjusting the intensity levels of a spotlight rather than turning it on and off to tune power levels up and down to match only the necessary processing activity.
Reducing the emphasis on the “fastest network”
The usual “fastest network” competition may no longer be relevant or meet MNO sustainability goals because 5G throughput is so high. MNOs that guarantee the fastest network services to their customers will always need to have all equipment power levels adjusted to the maximum – thus harming emissions control. Instead, it would be better for MNOs to focus on developing network slice SLAs and maintaining service levels at or slightly above each negotiated SLA given the enormous revenue potential of private 5G for businesses.
Due to the fact that 5G marks the first generation of mobile technology to include analytics in the framework, 5G standards are well suited for this kind of SLA monitoring. In order to manage multivendor networks, the 5G network data analytics function (NWDAF) provides standardized data collection. This data is easily accessible in real time.
Earlier mobile networks tended to use equipment from various vendors in different network domains, requiring separate analytics for each domain, which made comprehensive, predictive calculations all but impossible. Additionally, data was batch-processed, so there could be up to a 15-minute lag in availability, which is too long for enabling IoT and other real-time 5G applications.
Network slicing allows MNOs to manage power more precisely within their own networks as well as to participate in their customers’ sustainability programs for a fee. The operator can automate their macronetwork resource pool with granular power management by making aggregate, AI-driven forecasts about traffic loads with insight into each network slice. MNOs can predictively tune each enterprise customer’s private RAN/cell site to fulfil the customer’s power objectives and SLAs once a 5G base station has been assigned to them.
Turning it into reality
With carbon emissions a major concern and power expenses accounting for approximately 90% of network costs, there is a strong incentive to find solutions to curb 5G emissions. In accordance with the Paris Agreement, the mobile industry has jointly developed a climate plan of action to achieve net-zero emissions by 2050.
MNO executives should set goals for their firms to make sustainability a priority for every manager in order to comply and generate significant savings. Their programs should provide for the necessary training and regular evaluations of their performance.
In order to carry out their objectives, MNOs can follow the 5G NWDAF framework to deploy edge and core analytics. These systems can predict network demand in each cell and employ streaming analytics to evaluate, filter, and aggregate data in real-time as a way of reaching emissions reduction targets and cut costs.
The power consumption of their networks, i.e. the main contributor to carbon emissions, ought to be the primary focus of all operators. The most environmentally friendly companies are doing just that by implementing a thorough strategy that makes use of automation, AI, and predictive analytics tools. These technologies can provide the insights required to consistently achieve the optimal power-SLA balance to meet their energy-reduction objectives while providing the greatest 5G user experience.
