The intelligent cross-domain network operation and maintenance platform is aimed at provincial cloud networks and IP backbone networks. Based on large models and digital twin technology, it builds a full-process intelligent operation and maintenance capability of “pre-simulation prediction, in-process control, and decision support during incidents.” The platform integrates modules such as Wing Intelligence Distribution, Wing Simulation, Wing Scripting, and Wing Insight, achieving end-to-end fault visibility and digital employee closed-loop handling from configuration auditing and change simulation.
Currently, over 70% of network faults originate from human configuration errors. Traditional cutover processes heavily rely on manual reviews and night shifts, making it easy to overlook high-risk commands and error messages. The multi-vendor, multi-professional, and multi-level cloud network architecture makes business paths invisible and cross-professional collaboration difficult. The existing network data quality is inconsistent, tools are scattered, and there are copyright and security risks, making it challenging to support safe, efficient large-scale changes and rapid fault handling.
This project focuses on how telecom operators can leverage AIGC and network digital twin technology to drive digital transformation. The overall architecture of the project fully adopts the TM Forum open API and open data architecture (ODA) system, ensuring that the solution has long-term foresight and sustainability in terms of technological evolution, business interoperability, capability reuse, and ecological expansion. By integrating AIGC, automated scripts, secure change governance, cloud-network integrated visualization, and intelligent operation and maintenance capabilities, the project constructs an integrated end-to-end intelligent solution for the telecom industry’s complex network environment, fully reflecting the core advantages of automation, intelligence, and controllability.
The project solution consists of four major capability modules:① Wing Scripting / AI Script: Centered on large models and a structured template library, it achieves “full intelligent auditing + automatic optimization suggestions” through automatic identification of configuration syntax, dependencies, and potential risks, upgrading traditional manual sampling to a large model-driven configuration quality assurance system.
② Wing Simulation / AI Simulate: Constructs a highly realistic digital twin network to conduct pre-exercises for high-risk operations such as routing policies, cutover scripts, and new metropolitan area network reconstruction, ensuring that erroneous configurations “never enter the network,” achieving a critical leap from post-remediation to pre-validation.
③ Wing Intelligent Deployment / AI Deploy: Provides intelligent security control throughout the change process, including high-risk command identification, breakpoint verification, automatic shutdown on execution anomalies, real-time auditing, and backtracking, constructing a safe, controllable, and standardized change execution system to achieve efficient change management across provincial and municipal collaboration and cross-team cooperation.
④ Wing Insight / AI Insight: Integrates core networks, wireless, IPRAN, fixed broadband, transmission, and cloud resources to achieve a dynamic visualization of a unified topology and business links for cloud-network integration; supports “one event, one map” for fault perception, root cause analysis, and command scheduling, promoting a shift from manual multi-system comparisons to intelligent closed-loop decision-making.
This system forms an industry-leading full-process capability closed loop of “pre-simulation prediction — in-process control — intelligent decision-making during incidents,” significantly reducing operation and maintenance risks and providing a solid foundation for operators to create replicable, promotable, and sustainably evolving intelligent networks.
The project has been deeply applied in existing networks such as Guangdong Telecom, achieving 100% management of changes and business cutovers, reducing provincial and municipal collaboration time by 50%, and achieving a 100% standardization rate for changes; the accuracy rate of configuration auditing has improved to 93%-95%, covering over 34 core business scenarios and adapting to over 1200 configuration templates, significantly reducing operation and maintenance manpower and auditing costs. In a typical PoC, MTTR was reduced by 80%, and OPEX decreased by up to 40%, effectively avoiding large-scale network incidents and becoming a benchmark case for national cloud network operation and maintenance.
This project achieves a rare end-to-end intelligent operation and maintenance closed loop of “AI configuration auditing + high-precision simulation + intelligent script execution + cloud-network unified visualization,” with unified visibility and zero-code topology construction capabilities across manufacturers, professions, and levels; relying on the deep integration of large models and digital twins, it can be quickly replicated to other provinces and operators, and has already received the SNAI Best Case Award from the China Communications Association, as well as the provincial Science and Technology Progress First Prize and international recognition from TM Forum, highlighting its demonstrative and promotable value.
