Q&A: Tapping AI to ease the path to 5G
Steven Wu, VP of Huawei’s Carrier Business Group and President of Service & Software Marketing & Solution Sales Dept, explains how the company’s AI-based intelligent services can help carriers increase revenue, improve efficiency and accelerate the road to 5G.
Q: How does Huawei help operators accelerate their evolution to 5G?
Steven Wu (SW): Huawei started 5G research as early as 2009 – back when 4G was just starting to see commercial deployments. Over the past ten years, we invested more than $4 billion, which has enabled us to deliver end-to-end 5G equipment and services to operators.
The next-generation mobile service is important for operators because it offers the potential to help them attract more high-value subscribers and create new business opportunities across a wide variety of industries.
Our global service capabilities allow us to reduce time to market for operators. We have 30 years’ experience providing secure and reliable services, deploying sites in more than 170 countries and regions, and supporting stable operations of more than 1,500 network.
This experience is vital for accelerating end-to-end 5G implementations, which cover everything from planning and optimisation to network deployment and integration. Our professional services help operators maximise ROI and develop new services. Supporting our robust technologies, 22,000 frontline employees and more than 2,400 service suppliers are responsible for deploying solutions and providing support worldwide.
Huawei’s Digital Transformation Practice Centre is an incubator for third-party partners, providing them with a simulated 5G environment to develop, test and mature their business ideas.
Q: How can operators make use of existing 4G infrastructure to optimise 5G deployment?
SW: Huawei developed a tool to evaluate existing site facilities, tracking things like load capacity, to identify which sites can be reused for 5G. For sites that cannot be reused, Huawei has an innovative approach for fast and flexible deployment. For example, if a tower has insufficient load capacity, the approach improves installation efficiency 40 per cent by eliminating the need to drill and weld when upgrading the facility.
When deploying new 5G AAUs, Huawei considers the use of existing sites on the new network, then creates an antenna convergence reconstruction plan, which combines 2G, 3G and 4G antennas into one system. It then installs 5G AAUs in the space vacated, optimising the reconstructed network. The convergence plan ensures the original network performance does not deteriorate. In addition, more space is opened up for 5G AAU installation, reducing costs as less new space is required.
Q: What are the main challenges in 5G network planning?
SW: The high frequency of 5G spectrum delivers more limited coverage and weaker penetration compared with 4G bands. Huawei uses a special propagation model and 3D Massive MIMO simulation to improve planning accuracy. With the AI-based model, the simulation results in 10 per cent to 20 per cent better propagation. Huawei’s 3D simulation model can accurately position equipment not just to the outside of buildings, as most models can, but also inside facilities.
Different 5G services have different network requirements. In addition to providing a network simulation environment, Huawei also tests and pre-verifies specific use-cases in its OpenLab before commercial deployment, which can reduce the duration and risks of commercial delivery. In a rollout in South Korea we detected 90 per cent of key issues before commercial delivery.
Q: What is Huawei’s solution for reaching the best 5G throughout?
SW: Massive MIMO needs to perform 3D simulation and optimisation across as many as 10,000 patterns. This makes 5G optimisation about 100 times more complex than 4G. Huawei uses AI technology and algorithms to select the best site parameters in different scenarios.
Currently, most 5G networks are non-standalone. Huawei provides 4G and 5G synergetic optimisation to guarantee the end-user’s experience and maximise operators’ network value.
Q: Why is AI so important for implementing intelligent network operations?
SW: In the 5G era, new services, networks and technologies bring new challenges to operations. End-users have higher expectations. At the same time, services and networks become more complex and difficult to guarantee. Operations need to change from best-effort to offering clear and differentiated service level agreement (SLA) assurance.
This complexity has resulted in more outages caused by human error and rising operating costs. More than 70 per cent of network problems are caused by humans. It is necessary to shift operations using AI from reactive to predictive to improve the current low-quality and low-efficiency situation.
Q: Can you share some insight into Huawei’s progress in making systems proactive using learning-based tools?
SW: Huawei’s AUTIN (automation and intelligent) operations solution supports more than 20 predictive maintenance projects globally. Potential faults and risks can be predicted up to 30 days in advance, giving operators ample time to make the necessary corrective measures.
We leverage our assets and experience to offer a robust predictive maintenance service. Our vision is to ultimately reach close to zero network outages. We have achieved prediction capabilities addressing use-cases on service and hardware faults in specific scenarios. In the coming years we will be adding more use-cases as well as extending it to 5G networks.
Q: What do you see as the biggest obstacle in slowing operators’ efforts to improve their network operations?
SW: With the launch of 5G, operators operate in a complex environment with 2G, 3G, 4G and 5G networks co-existing. The number of managed objects will increase significantly, which will generate more alarms and work orders each day. The requirements for SLA and customer experience assurance will change as well, with even higher demands on capacity, availability and quality. Although operators have 21st century technology across 4G and 5G networks, many of them are still using the traditional way to manage in this complicated environment.
The AUTIN solution has accumulated a large number of operation assets which enabled us to improve operation quality and efficiency. Since early 2019, we have focused on forecasting and preventing potential problems in wireless devices using AI algorithms and developing commercial assurance solutions for network slicing. In future, we will look to develop self-healing, self-optimisation capabilities for 5G networks and further improve the assurance capability of the network slicing service.
Q: How can operators speed up 5G site rollouts?
SW: Time to market is important for operators as an early start can help them attract high-value subscribers. However, building a 5G network is not easy work. To speed up the deployment of a high-quality 5G network, Huawei offers an AI-based digital solution for site installation.
The first step is to create a 3D digital twin of an actual site after engineers complete a detailed site survey. This speeds up site design as measurements can first be done on the digital twin, with engineers correcting site faults online and eliminating the need for frequent site visits. Huawei is able to reduce rollout times by 30 per cent. For example, 3,000 sites were deployed in South Korea within a month.
Q: Why is there so much talk about 5G energy usage with 5G?
SW: The power consumption of 5G per bit is about 1/50 of 4G. That’s a huge reduction. However, 5G uses much more spectrum, resulting in energy consumption significantly increasing to achieve the same coverage. Energy savings has become an important consideration for operators.
Through continuous improvement in chips, power amplifiers and algorithms, Huawei 5G gear uses 20 per cent less energy than the industry average. In addition, with AI online energy savings, Huawei’s PowerStar can significantly reduce the power consumption of telecoms equipment.
“Although operators have 21st century technology across 4G and 5G networks, many of them are still using the traditional way to manage in this complicated environment.”