A Scalable Rule-Based and AI-Assisted Framework for Archiving in Large-Scale Telecom Catalog Systems

This study presents a rule-based and AI-supported archiving strategy developed to address the data management and performance challenges observed in Turkcell’s Central Product Catalog Management (CPCM) system. As the volume of data has steadily increased, the system has begun to face critical issues related to performance and long-term sustainability. Traditional archiving techniques were deemed inadequate due to the system’s versioning architecture and strict regulatory constraints. To overcome these limitations, we propose a hybrid framework consisting of two stages: (1) rule-based pre-archiving defined by domain expertise, (2) AI-based access analysis to identify frequently accessed data to ensure scalability and operational continuity. Initial evaluations indicate significant archiving potential: up to 51% for EPOS offers and 16% for campaign data. Furthermore, the framework aims to reduce average query response times from 600–750 ms to 300–400 ms without requiring structural changes to the monolithic system. This approach offers a practical solution for sustainable data management in large-scale telecommunications systems and contributes to academic discussions on hybrid archiving strategies, performance optimization, and regulatory compliance in enterprise information systems.