MCP for Network Management: Problem Statement, Use Cases, and Requirements

The emergence of large language models (LLMs) and AI agents is reshaping how network operators interact with infrastructure. However, current network management systems lack a standardized, secure, and intent-driven interface that enables AI agents to discover, invoke, and reason over network capabilities. This document presents a problem statement for integrating the Model Context Protocol (MCP) into network management, outlines key use cases—including troubleshooting, measurement, security, and optimization—and specifies functional, security, and interoperability requirements for MCP-based network management architectures.