Self-Healing in Knowledge-Driven Autonomous Networks: Context, Challenges, and Future Directions

With the advancement of communication and computer technology, network architectures have become increasingly complex, accompanied by the continual emergence of novel services. This increase in sophistication surpasses the abilities of manual maintenance and traditional network management solutions, thereby elevating the potential risk of network failures. This article provides a comprehensive review of significant network incidents that have occurred over the past decade and outlines the evolution of network self-healing. Building on these findings, we propose an effective knowledge-driven self-healing pattern (KSHP) suitable for a broad spectrum of network failure scenarios, designed to alleviate the challenges of automation and scalability in autonomous networks. In our case study, we implemented a self-healing agent following the KSHP schema to mitigate network service failures caused by link congestion. The experimental results validate the efficacy and practicality of KSHP in network fault scenarios. Furthermore, future directions are discussed and analyzed in detail. The insights presented in this article provide a basic pattern for the knowledge-driven self-healing framework, facilitating the achievement of Level 5 autonomous networks.