Concept & Methodology
Agentic6G Overall Concept
Future Beyond-5G and 6G networks are expected to evolve into AI-native infrastructures capable of supporting highly dynamic services and intelligent automation across the device–edge–cloud continuum. These environments will require autonomous and adaptive mechanisms for service orchestration, lifecycle management, and distributed decision-making, enabling networks to respond in real time to changing operational conditions, service demands, and application requirements.
The Agentic6G approach is inspired by the concept of distributed Agentic AI for future communication networks, aiming to introduce autonomous multi-agent intelligence into AI-native 6G ecosystems. The project envisions a collaborative environment where distributed intelligent agents are capable of coordinating actions, reasoning on operational objectives, and dynamically adapting network and service behaviour across heterogeneous infrastructures.
Agentic6G aspires to establish a trustworthy and scalable AgenticOps framework that supports autonomous service management, intelligent orchestration, and resilient operation of distributed AI-native services and network functions. By combining knowledge-driven reasoning, hybrid orchestration, and self-organising capabilities, the project contributes to enabling scalable distributed intelligence for future Beyond-5G and 6G environments.
The Project Execution Pillars
1
Agentic AI Framework
An Agentic AI framework enabling autonomous collaboration and intelligent coordination between distributed agents across device-edge-cloud infrastructures.
2
Hybrid Orchestration & Lifecycle Management
A hybrid orchestration and lifecycle management framework supporting adaptive deployment, monitoring, scaling, and self-healing of intelligent services and network functions.
3
Trustworthy & Resilient AI-Native Operations
Trustworthy and resilient AI-native operational mechanisms integrating security, resilience, interoperability, and operational intelligence into distributed multi-agent environments.
4
Validation & Demonstration
Validation and demonstration activities across representative Beyond-5G/6G scenarios and heterogeneous infrastructures to evaluate scalability, adaptability, and operational efficiency.
Agentic6G Scientific Methodology
Agentic6G aims to deliver a scalable and trustworthy multi-agent framework capable of supporting autonomous AI-native operation in future Beyond-5G and 6G communication environments.
The project methodology follows three iterative and interconnected phases designed to progressively design, implement, integrate, and validate the Agentic6G framework and its associated technologies.
Project Execution Pillars
Phase 1
Requirements Analysis and System Design
The first phase focuses on the analysis of project use cases, stakeholder requirements, and operational challenges related to AI-native 6G ecosystems. Functional and technical requirements are translated into system specifications and architectural designs supporting distributed multi-agent orchestration, service lifecycle management, security, resilience, and interoperability. This phase also defines the target Key Performance Indicators (KPIs) and validation requirements that will guide the implementation and evaluation activities.
Phase 2
Framework Implementation, Integration and Orchestration
The second phase focuses on the implementation of the Agentic AI framework, orchestration mechanisms, operational intelligence components, and lifecycle management capabilities. The project integrates distributed agents, hybrid orchestration approaches, knowledge-driven reasoning mechanisms, and self-organising functionalities such as spontaneous agent generation, self-healing, and auto-scaling.
The implementation activities follow iterative cycles of integration and validation, enabling continuous refinement of the framework components and orchestration mechanisms across heterogeneous device-edge-cloud environments.
Phase 3
Validation and Demonstration
The final phase focuses on the large-scale validation and demonstration of the Agentic6G framework across representative Beyond-5G and 6G scenarios. Validation activities evaluate the scalability, resilience, trustworthiness, orchestration efficiency, and operational adaptability of the proposed multi-agent framework under realistic and dynamic operating conditions.
Iterative validation and feedback cycles will enable continuous optimisation of the framework and operational mechanisms, ensuring progressive refinement of the Agentic AI capabilities and supporting the maturity of the proposed solutions for future AI-native 6G ecosystems.