Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ORION: Intent-Aware Orchestration in Open RAN for SLA-Driven Network Management

Gabriela da Silva Machado, Gustavo Z. Bruno, Alexandre Huff, Jose Marcos Camara Brito, Cristiano B. Both

TL;DR

ORION is proposed, an O-RAN compliant intent orchestration framework that integrates Large Language Models via the Model Context Protocol (MCP) to translate natural language intents into enforceable network policies, paving the way for autonomous 6G networks.

Abstract

The disaggregation of the Radio Access Network (RAN) introduces unprecedented flexibility but significant operational complexity, necessitating automated management frameworks. However, current Open RAN (O-RAN) orchestration relies on fragmented manual policies, lacking end-to-end intent assurance from high-level requirements to low-level configurations. In this paper, we propose ORION, an O-RAN compliant intent orchestration framework that integrates Large Language Models (LLMs) via the Model Context Protocol (MCP) to translate natural language intents into enforceable network policies. ORION leverages a hierarchical agent architecture, combining an MCP-based Service Management and Orchestration (SMO) layer for semantic translation with a Non-Real-Time RIC rApp and Near-Real-Time RIC xApp for closed-loop enforcement. Extensive evaluations using GPT-5, Gemini 3 Pro, and Claude Opus demonstrate a 100% policy generation success rate for high-capacity models, highlighting significant trade-offs in reasoning efficiency. We show that ORION reduces provisioning complexity by automating the complete intent lifecycle, from ingestion to E2-level enforcement, paving the way for autonomous 6G networks.

ORION: Intent-Aware Orchestration in Open RAN for SLA-Driven Network Management

TL;DR

ORION is proposed, an O-RAN compliant intent orchestration framework that integrates Large Language Models via the Model Context Protocol (MCP) to translate natural language intents into enforceable network policies, paving the way for autonomous 6G networks.

Abstract

The disaggregation of the Radio Access Network (RAN) introduces unprecedented flexibility but significant operational complexity, necessitating automated management frameworks. However, current Open RAN (O-RAN) orchestration relies on fragmented manual policies, lacking end-to-end intent assurance from high-level requirements to low-level configurations. In this paper, we propose ORION, an O-RAN compliant intent orchestration framework that integrates Large Language Models (LLMs) via the Model Context Protocol (MCP) to translate natural language intents into enforceable network policies. ORION leverages a hierarchical agent architecture, combining an MCP-based Service Management and Orchestration (SMO) layer for semantic translation with a Non-Real-Time RIC rApp and Near-Real-Time RIC xApp for closed-loop enforcement. Extensive evaluations using GPT-5, Gemini 3 Pro, and Claude Opus demonstrate a 100% policy generation success rate for high-capacity models, highlighting significant trade-offs in reasoning efficiency. We show that ORION reduces provisioning complexity by automating the complete intent lifecycle, from ingestion to E2-level enforcement, paving the way for autonomous 6G networks.
Paper Structure (36 sections, 1 equation, 12 figures, 2 tables)

This paper contains 36 sections, 1 equation, 12 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background
  3. O-RAN Architectural Elements
  4. Policies vs. Intents
  5. Intent Lifecycle
  6. Roles in Intent-Driven RAN
  7. Model Context Protocol
  8. Related Work
  9. Proposed Architecture: ORION
  10. Design Objectives
  11. High-Level Architecture Overview
  12. Component Details
  13. End-to-End Intent Workflow
  14. Extensibility and Standards Alignment
  15. Implementation and Prototype
  16. ...and 21 more sections

Figures (12)

  • Figure 1: Intent lifecycle states and transitions with default flow (1--6) and CLA feedback loops (7--8).
  • Figure 2: ORION high-level architecture spanning SMO/Non-RT and Near-RT RIC: the MCP Client/Server structure intents using CAMARA's NetworkSliceBooking schema; the OrApp composes policies, and the OxApp enforces them via A1/E2.
  • Figure 3: Non-RT RIC Intent Execution Flow.
  • Figure 4: Sequence diagram of the control loop.
  • Figure 5: Token consumption per intent across LLM providers (log scale). Solid bars represent input tokens; hatched bars represent output tokens. Error bars indicate standard deviation.
  • ...and 7 more figures