Scholarlink

<p>Large language models (LLMs) are increasingly explored to support telecom data management across OSS/BSS operations, network automation, and telemetry-driven analytics. This systematic review examines how LLMs are applied to implement and enhance DAMA-DMBOK knowledge areas in telecommunication systems. Following a PRISMA 2020 process, we searched 10 academic and industry/standards sources (2021–January&#xa0;2026) and screened 1,062 records, resulting in 11 peer-reviewed studies (2023–2025). We map each study to DAMA domains and characterize the LLM mechanism and evaluation setting. Our findings show that studies mostly concentrated on Data Integration &amp; Interoperability (55%) and Metadata Management (36%), with limited coverage of Data Governance (27%), Data Quality (18%), Data Security (9%), Data Architecture (18%), and Data Warehousing &amp; Business Intelligence (18%). Data Modeling &amp; Design and Data Storage &amp; Operations remain unexplored in the included evidence base. Retrieval-augmented generation is the most common mechanism (45%), followed by agentic/tool-using systems (36%) and fine-tuning (18%); only 27% of studies report latency despite carrier-grade operational constraints. Most evaluations are non-production (82%), limiting generalizability and leaving multi-vendor interoperability and compliance enforcement weakly evidenced. We synthesize cross-cutting patterns and outline research directions, including telecom-aligned benchmarks, standards-driven evaluation, and longitudinal production-scale studies to close the gap between promising prototypes and operator-grade deployment.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Large language models for DAMA-aligned data management in telecommunication systems: a review of OSS/BSS and 5G/6G applications

  • 期刊论文

  • 发表日期：

    2026年5月28日

查看全文
  • Omar I. Alsaleh

摘要

Large language models (LLMs) are increasingly explored to support telecom data management across OSS/BSS operations, network automation, and telemetry-driven analytics. This systematic review examines how LLMs are applied to implement and enhance DAMA-DMBOK knowledge areas in telecommunication systems. Following a PRISMA 2020 process, we searched 10 academic and industry/standards sources (2021–January 2026) and screened 1,062 records, resulting in 11 peer-reviewed studies (2023–2025). We map each study to DAMA domains and characterize the LLM mechanism and evaluation setting. Our findings show that studies mostly concentrated on Data Integration & Interoperability (55%) and Metadata Management (36%), with limited coverage of Data Governance (27%), Data Quality (18%), Data Security (9%), Data Architecture (18%), and Data Warehousing & Business Intelligence (18%). Data Modeling & Design and Data Storage & Operations remain unexplored in the included evidence base. Retrieval-augmented generation is the most common mechanism (45%), followed by agentic/tool-using systems (36%) and fine-tuning (18%); only 27% of studies report latency despite carrier-grade operational constraints. Most evaluations are non-production (82%), limiting generalizability and leaving multi-vendor interoperability and compliance enforcement weakly evidenced. We synthesize cross-cutting patterns and outline research directions, including telecom-aligned benchmarks, standards-driven evaluation, and longitudinal production-scale studies to close the gap between promising prototypes and operator-grade deployment.