Autonomous Network Construction: E2E Automatic Collaboration for Complaint Handling and Hidden Risk Identification

As 5G networks are extensively deployed and the digital economy continues to grow, the construction and operation of communication networks face more requirements and challenges. In particular, the proliferation of multiple frequencies and standards has led to increasingly complex and vast communication network structures. Diverse terminals are heavily connected, accompanied by a rapid surge in different types of service requirements. Traditional O&M methods are no longer adequate to meet customer efficiency and cost expectations, nor can they fulfill the demands for user experience. With automation and intelligence as its core features to improve network quality and efficiency, an autonomous network is becoming a crucial trend in the future of communication networks.

In 2017, ETSI established the first standard organization for network intelligence, and later, ITU/3GPP/CCSA formed intelligent network projects. The development of autonomous networks began to take shape in 2019. Though certain areas have achieved partial automation capabilities, they mainly concentrate on intelligent tools and functionalities for individual domains and scenarios, lacking complete end-to-end automation. As a result, interconnections between different fields rely on manual efforts, lacking automatic collaboration.

As the industry’s leading provider of integrated communications and information technology solutions, ZTE has conducted ongoing research into automation capabilities in areas like wireless, bearer, and core networks. It has achieved industry-leading levels in certain single-domain autonomous functionalities. However, to enhance autonomous capabilities of communication networks and assist operators in their digital transformation, it is necessary to improve end-to-end (E2E) autonomous capabilities. Therefore, ZTE has proposed a full-domain autonomous network solution leveraging its VMAX big data platform. The solution seamlessly integrates diverse single-domain capabilities to achieve end-to-end intelligence. Particularly, two end-to-end scenarios—customer complaint handling and hidden risk identification—have been successfully implemented in multiple projects, thus facilitating operators’ network intelligence enhancement and driving digital transformation.

ZTE’s uSmartNet full-domain autonomous network solution, built on its big data product VMAX, establishes seamless connectivity between various individual domains through relevant standard interfaces. It retrieves relevant configuration, alarm, and performance data from wireless, bearer, and core networks. Through the analysis and processing of big data, it intelligently analyzes and precisely locates problems, ranging from individual call detail records to the entire network. Moreover, VMAX extends its capabilities and interconnects intelligent tools from different domains through the OpenAPI interface. Leveraging intelligent core algorithms, this approach addresses the challenge of diverse protocols and characteristics in individual domains, which previously hindered collaborative efforts. By taking a multi-layered approach that encompasses native network elements, single-domain automation, and cross-domain collaboration, it enables automatic closed-loop of full-domain network automation and problem handling (Fig. 1).

In the complaint handling scenario, the typical approach heavily relies on manual intervention with tools from individual domains as auxiliary support. However, this can result in lengthy complaint processing cycles and substantial expenditure of manpower resources. To address these challenges, ZTE leverages its VMAX big data platform that employs cutting-edge data mining algorithms to achieve end-to-end automatic complaint demarcation, covering terminals, transmission, wireless, core networks, and SPs. The platform also correlates wireless measurement reports (MR), performance data, and alarm data. Through fault tree analysis across various complaint user scenarios, user-level root causes can be drilled, and problems can be located and handled quickly. In some scenarios, automatic closed-loop of complaint handling can be achieved. For example, in the TNR coverage problem analysis, the platform utilizes ZTE’s wireless intelligence function—automatic antenna pattern control (AAPC) through an OpenAPI interface. By employing the AAPC weight optimization solution, it automatically distinguishes the scenarios with coverage problems and establishes a relative coordinate system based on the direction of arrival (DOA) data, enabling automatic weight optimization without relying on engineering parameters. This ensures personalized weight deployment and automatic optimization with a “one-station, one-scenario, one-weight” approach, leading to true “zero-touch” fully automated optimization. According to the evaluation criteria of TMF Autonomous Network Whitepaper 4.0, this scenario almost achieves L4 automation capability. Additionally, it interconnects the BN intelligence tool BigDNA and the CN intelligent tool Magiceye, enabling automatic demarcation and location of bearer and core network problems. Furthermore, the interconnection with wireless intelligence tool AAX facilitates precise locating of wireless-side alarm problems and software self-healing, ultimately achieving end-to-end automatic handling of complaint problems.

In the hidden risk identification scenario, the industry is still in the early stage of exploration, with most network elements being handled only after faults occur, resulting in fluctuations in network quality that affect user experience. To address this issue, ZTE utilizes its OSS intelligent tool AIMIND, which accesses CM/PM/FM data from wireless, bearer, and core networks, and employs AI algorithms and self-learning capabilities to forecast potential site outages. According to the forecast results, operators can take proactive measures to mitigate the risk of site outages, ensuring the stable operation of communication networks. Moreover, AIMIND provides targeted fault predictions in various individual domains, including wireless-side optical module, RRU/AAU environmental issues, voltage-related problems, as well as bearer hardware and traffic-related faults. Its fault prediction capability stands at the forefront of the industry.

ZTE’s full-domain autonomous network construction solution uSmartNet has been implemented in multiple projects. Particularly, in the customer complaint handling scenario, it has achieved true end-to-end automatic network optimization. In some scenarios, the solution accomplishes automatic closed-loop of problem handling without the need for manual intervention. This remarkable approach significantly reduces complaint handling time and minimizes the need for expert involvement, thereby facilitating operators’ digital transformation. Moreover, in the hidden risk identification scenario, the solution automatically identifies hidden risks and gives suggestions. This capability helps operators improve fault handling efficiency by over 30% and effectively reduces the probability of fault occurrence. As a result, it enhances network stability and enables industry-leading intelligent network O&M, empowering operators to stay ahead in the market. In addition to achieving automation capabilities in the above-mentioned scenarios, ZTE has also deployed end-to-end solutions for FNR coverage and wireless interference issues, expanding the scope of operators’ autonomous networks.

Full-domain automation is a vital path to the development of communication intelligence. Through intelligent technologies, it enables cross-domain collaborative intelligent network O&M, encompassing wireless networks, bearer networks, and core networks. This advancement leads to improved operational efficiency and enhances network reliability, security, and user experience.