Huawei is accelerating the development of agentic networks, aiming to create an infrastructure where AI agents manage network connections.

Agentic Core and A2A-T

Huawei has introduced Agentic Core, a three-engine network architecture designed to support the needs of AI agents. Additionally, it promoted A2A-T (Agent-to-Agent for Telecom), an open-source protocol for collaboration between agents in telecommunications environments, developed in collaboration with industry partners and formalized by TM Forum.

The Interoperability Challenge

Currently, network architectures are optimized for traffic generated by human users. AI agents, with their specific bandwidth and latency requirements, demand a different approach. Huawei proposes "intent-driven" networks, where AI within the network interprets requests and dynamically allocates resources.

Interoperability between agents is crucial. Google launched the Agent2Agent (A2A) protocol for enterprise-level integration. Huawei's A2A-T, on the other hand, focuses on the network layer, addressing reliability, security, and latency requirements specific to telecommunications. Without a shared protocol, every interaction between agents would require custom integrations, limiting scalability.

Open Source for Adoption

Huawei has chosen to open source A2A-T to accelerate its adoption. The project includes an SDK for standardized agent integration, a Registry Center for authentication and skills management, and an Orchestration Center with low-code/no-code visual tools. The goal is to lower the barriers to entry for operators and integrators, simplifying the implementation of agentic networks.

For those evaluating on-premise deployments, there are trade-offs to consider. AI-RADAR offers analytical frameworks on /llm-onpremise to evaluate these aspects.