DOI: https://doi.org/10.36347/sjet.2026.v14i05.003

Pages: 208-221

Long distance fiber communication networks are a crucial element of modern digital infrastructure. They are used for broadband access, cloud computing, business connectivity, as well as for data transport over a wide geographic area. However, reliability issues are related to signal degradation, component failure, alarm repetition, service fluctuation, as well as maintenance delays in a distributed environment. Conventional monitoring methods often depend on alarm reporting and post-failure response. Such approaches provide limited visibility into gradual degradation and emerging faults. This paper proposes a generic reliability monitoring framework for long-distance fiber communication networks to address this gap. The framework combines link-level, node-level, service-level, and maintenance-history indicators within a unified analytical structure. It applies parameter normalization, a Reliability Monitoring Index for condition assessment, and a Fault Risk Score for maintenance prioritization. The methodology includes data acquisition, preprocessing, feature construction, reliability-state classification, and operator-oriented output interpretation. The discussion shows that the framework can represent gradual degradation, recurring instability, and service-critical faults in a clear and structured form. This study provides a conceptual foundation for continuous reliability monitoring, early fault identification, and maintenance planning in long-distance fiber communication systems. It also offers a basis for future validation with simulated datasets and operational network records.