RAN Intelligence And The Pursuit Of 5G Energy-efficiency

Sponsored Feature As rollouts of 5th generation (5G) mobile networks gain momentum worldwide, telcos and mobile network operators (MNOs) are putting substantial capital into their infrastructure. And inevitably, the pressure to recoup the vast funds being invested is likely to intensify in parallel with increased service demands and operational costs.

Today, energy costs represent between 20 percent and 40 percent of a telco's operational expenditure (opex), according to the GSMA. It predicts that network traffic will grow threefold by 2025 compared with 2021 as the number of global 5G connections hits 2.2 billion. But on the other side of the coin, its expected that 5G networks handling a substantial increase in data traffic could arguably double or triple operators' energy costs when it comes to preserving network quality and optimizing the end user experience.

ZTE is among a growing number of companies seeking ways to help MNOs reduce that opex by giving them the means to maintain sustainable operations. In a GSMA network transformation survey, 52 percent of operators rated "service operations automation" as the most promising initiative for opex reduction.

This is not surprising considering the complex networks which now feature a mixture of 2G/3G/4G/5G multi-band and multi-mode technologies that MNOs now have to grapple with, most of which are also required to support diverse services such as augmented and virtual reality (AR/VR), the Internet of Things (IoT), autopilot, machine vision and holographic calls. Contributing further to those demands is the use of three new service attributes: Ultra-Reliable and Low Latency Communications (URLLC); Enhanced Mobile Broadband (eMBB); and Massive Machine-Type Communications (mMTC).

"Operators are now facing challenges to improve the user experience and enhance utilization of expensive, yet limited, radio frequency spectrum to stay competitive, while the old-fashioned 'one-fits-all' network policy never meets individual needs and always consumes more energy and resources than needed," said Ms Tang Xue, Vice President of ZTE's RAN Product Line. "Radio access network (RAN) intelligence is a smart way to get more juice out of the already deployed infrastructure."

RAN intelligence can play a pivotal role in enabling more process automation, service optimization and energy efficiency in telco infrastructure. Based on network and base station-native artificial intelligence (AI) technology, ZTE RAN Intelligence can be extended from smart operations and maintenance (O&M) to intelligent service guarantees and power saving.

Intelligent orchestration drives infrastructure efficiency

ZTE's Radio Composer intelligent orchestration solution for example uses automatic service recognition and traffic pattern learning to enable precise mapping of a variety of service requirements to network resources and terminal capabilities. They make network infrastructure utilization more efficient. It has been deployed in more than 10,000 sites with commercial trials underway in Thailand, the Philippines and other markets.

These trials help operators to draw more traffic as they improve user experience and service reliability, says ZTE. Ms Tang states: "According to the trial result, Radio Composer increases network payload by 30 percent without additional spectrum, which stimulates new business plans, improves data package sales, and facilitates new service applications like Extended Reality and Vehicle-to-Everything."

Instead of requiring MNOs to roll out more sites or invest in new spectrums, Radio Composer leverages the native AI inside base stations to enable local data processing and relax bandwidth requirements for raw data transmission. This reduces operators' transmission power consumption, says ZTE. So for a network with thousands of base stations, Radio Composer can save substantial capital expenditure (capex) as well as O&M costs.

This benefit is further boosted by computing power sharing. "Computing power orchestration prepares the upper layer applications with needed intelligence," Ms Tang explained. "It coordinates computing power resources among a range of base stations and pools extra computing power from idle sites to busy sites, significantly increasing power efficiency. Such orchestration can increase available computing power supporting busy sites by 14 percent, significantly saving investments in external power sources."

The commercial trials also aim to cut operators' opex, according to ZTE. Radio Composer implements intent-driven networking to automatically guarantee the user experience, for example. By using natural language to communicate an intent, such as "TikTok user experience at the stadium with top priority", the system can process the service guarantee flow automatically to prioritize that application. Self-generated scheduling policies, performance feedback, and automatic optimization of the service experience can all help to reduce the telecommunications engineering manpower needed to maintain those operations.

Radio Composer trials in China

ZTE has already worked with China Mobile to enhanced the MNO's user experience, network traffic and network efficiency on its existing infrastructure in China using Radio Composer. In Hangzhou, the two companies simulated the spatial location flow characteristics of user groups in a large venue. At different times during big events such as concerts and exhibitions, a Radio Composer-driven solution helped to predict the traffic distribution and to support intelligent user scheduling so the network could cope with high network load and differentiated user requirements.

Elsewhere computing power orchestration logically aggregated multiple 5G base stations to create a power resource pool, allowing a base station node with remaining power in the pool to enable another base station to perform data analysis and AI calculations during busy hours. The results were encouraging. WeChat pictures and videos were successfully sent and webpages opened within 3 milliseconds (ms). OTT video calls and live streaming were clear and free of lag, while voice-over-NR voice calls were fully connected, reported ZTE and China Mobile.

In Hebei province, a 5G cloud desktop project at Cangzhou Central Hospital leveraged a NodeEngine computing board in a 5G base station geared for B2B edge computing and Radio Composer in the baseband unit for B2C service. The solution dynamically identified service flow types and quickly split and established quality-of-service (QoS) flows to deliver accurate service guarantees and circumvent the complicated contract management process associated with user SIM cards. The intelligent allocation of computing power on demand also allowed the mobile network to split data stream for latency-sensitive services from the default data stream. This reduced the average latency of service data from 16ms to about 10ms, said the company.

Elsewhere in Fujian, an intent-driven Smart RAN commercial trial allocated network resources according to service priorities. It covered more than 7,000 identifiable service types covering 95 percent of mainstream mobile services relating to common applications including epidemic prevention code scanning, TikTok live streaming, Tencent video streaming and WeChat online conferencing. "The trial in Fujian with China Mobile has reduced O&M cost to one-fifth of original, with improved user throughput as well as decreased latency." Ms Tang says.

The contribution of ZTE's Radio Composer to helping MNOs in China realize these benefits led to the company winning the "Innovative Mobile Service and Application Award (Outstanding Award)" at the 2023 Global TD-LTE Initiative (GTI) Summit as well as the "Best Technology Deployment" award at the 2022 Telecom Review Leaders' Summit.

Zero-load, zero-carbon reduce base station power consumption

Beyond user and power orchestration, "zero-load, zero-carbon" goals have led ZTE to lower the power consumption of 5G mobile base stations. In general, a 5G macro base station consists of one baseband unit (BBU) and three Active Antenna Units (AAUs), so embedding power saving technologies in these components can make a real difference in helping telcos and MNOs future-proof their investment in greener technology, says ZTE.

"With cutting-edge technologies and advanced chipsets, the power consumption of new 5G massive-Multiple-Input Multiple-Output (MIMO) products has been reduced by 35 percent," Ms Tang reckons. "The innovative power amplifier (PA) architecture in RRU platforms also increases the PA efficiency for 4G modernization. Besides, ZTE has been adopting new materials and structural design for better heat dissipation."

To the same end, ZTE has developed a shutdown mechanism called AAU Hibernation, which is yet another energy saving innovation that helps save even more energy than the existing features such as deep sleep, symbol shutdown and channel shutdown. This mechanism can cut an AAU's residual power consumption to less than 5W in zero-load state, potentially saving 6.25 million kWh per 10,000 stations annually, says the company.

AAU Hibernation is a key feature of ZTE's PowerPilot Pro solution which combines with Pico Remote Radio Unit (pRRU) auto-sensing and base station-native AI and network AI functions to reduce RAN energy consumption. ZTE's PowerPilot Pro has already been deployed in more than 30 countries and in more than 900,000 base stations. The Sichuan branch of China Telecom for example has rolled out PowerPilot Pro's AAU hibernation across Leshan city while digital indoor distribution with pRRU auto-sensing in Chengdu city has also helped reduce the 5G networks' carbon footprint.

Indeed, the company expects the AAU hibernation feature to save 7 million kWh every year, equivalent to a 4,200 ton reduction in CO2 emissions. The deployment of pRRU auto-sensing in the digital indoor distribution systems is also expected to cut energy consumption by 25 percent.

Additionally, ZTE has introduced its uSmart RNIA intelligent wireless network solution which features network analysis, control and management capabilities. This is aimed at helping operators build networks simply, quickly and efficiently at low cost by applying AI across the development stages of the network life cycle – everything from planning, deployment, maintenance to optimization and operation scenarios.

ZTE's continued quest for greener 5G technologies shows that energy savings are more likely to be achieved by addressing many factors simultaneously. And the pressure on global operators to lower carbon emission will invariably require network infrastructure to be increasingly intelligent.

Sponsored by ZTE.