SANNet: A Semantic-Aware Agentic AI Networking Framework for Multi-Agent Cross-Layer Coordination
Yong Xiao, Haoran Zhou, Xubo Li, Yayu Gao, Guangming Shi, Ping Zhang
TL;DR
This work tackles the challenge of coordinating a large set of heterogeneous AI agents across application-, physical-, and network-layers to fulfill user semantic goals in real-time networks. It introduces SANNet, a semantic-aware agentic AI networking architecture featuring an agent controller that detects user goals, decomposes them into layer-specific subtasks, and dynamically selects and mediates participating agents. The authors formalize the multi-agent cross-layer problem, define C-error and G-error as metrics of conflict and generalization, and propose a dynamic weighting scheme to steer the system toward Pareto-optimal behavior, with theoretical bounds and guarantees. A hardware prototype on open RAN and 5G core validates the approach, showing substantial reductions in conflict (C-error) and demonstrating feasible semantic-goal-driven orchestration in a realistic environment, signaling a path toward autonomous, self-optimizing networks.
Abstract
Agentic AI networking (AgentNet) is a novel AI-native networking paradigm that relies on a large number of specialized AI agents to collaborate and coordinate for autonomous decision-making, dynamic environmental adaptation, and complex goal achievement. It has the potential to facilitate real-time network management alongside capabilities for self-configuration, self-optimization, and self-adaptation across diverse and complex networking environments, laying the foundation for fully autonomous networking systems in the future. Despite its promise, AgentNet is still in the early stage of development, and there still lacks an effective networking framework to support automatic goal discovery and multi-agent self-orchestration and task assignment. This paper proposes SANNet, a novel semantic-aware agentic AI networking architecture that can infer the semantic goal of the user and automatically assign agents associated with different layers of a mobile system to fulfill the inferred goal. Motivated by the fact that one of the major challenges in AgentNet is that different agents may have different and even conflicting objectives when collaborating for certain goals, we introduce a dynamic weighting-based conflict-resolving mechanism to address this issue. We prove that SANNet can provide theoretical guarantee in both conflict-resolving and model generalization performance for multi-agent collaboration in dynamic environment. We develop a hardware prototype of SANNet based on the open RAN and 5GS core platform. Our experimental results show that SANNet can significantly improve the performance of multi-agent networking systems, even when agents with conflicting objectives are selected to collaborate for the same goal.