Efficient Asynchronous Federated Evaluation with Strategy Similarity Awareness for Intent-Based Networking in Industrial Internet of Things

TL;DR

This work tackles the challenge of safe, scalable verification and deployment of intent-based policies in heterogeneous IIoT environments with privacy constraints. It introduces FEIBN, a Federated Evaluation Enhanced IBN framework that uses multimodal intent alignment with LLMs to translate user intents into structured strategy tuples and applies federated evaluation for distributed policy verification. To address training efficiency and communication overhead, it presents SSAFL, a Strategy Similarity Aware Federated Learning mechanism with similarity-based node selection and magnitude-based asynchronous updates. Experimental results show that SSAFL improves accuracy and convergence speed while significantly reducing communication, demonstrating practical scalability for industrial deployments.

Abstract

Intent-Based Networking (IBN) offers a promising paradigm for intelligent and automated network control in Industrial Internet of Things (IIoT) environments by translating high-level user intents into executable network strategies. However, frequent strategy deployment and rollback are impractical in real-world IIoT systems due to tightly coupled workflows and high downtime costs, while the heterogeneity and privacy constraints of IIoT nodes further complicate centralized policy verification. To address these challenges, we propose FEIBN, a Federated Evaluation Enhanced Intent-Based Networking framework. FEIBN leverages large language models (LLMs) to align multimodal user intents into structured strategy tuples and employs federated learning to perform distributed policy verification across IIoT nodes without exposing raw data. To improve training efficiency and reduce communication overhead, we design SSAFL, a Strategy Similarity Aware Federated Learning mechanism that selects task-relevant nodes based on strategy similarity and resource status, and triggers asynchronous model uploads only when updates are significant. Experiments demonstrate that SSAFL can improve model accuracy, accelerate model convergence, and reduce the cost by 27.8% compared with SemiAsyn.

Figures (7)

• Figure 1: The FEIBN framework for IIoT. The framework supports intent-driven strategy generation and verification across distributed IIoT nodes. Within this framework, an LLM translates multimodal user intents into strategy tuples, and asynchronous federated policy verification is performed based on a similarity-aware node selection mechanism.
• Figure 2: The process of SSAFL. Nodes first evaluate their strategy similarity and resource availability, and only those satisfying the selection criteria join the current round. Selected nodes receive the global model, perform local training, and compute adaptability scores. Nodes whose adaptability meets the upload threshold send updates asynchronously to the server. The server aggregates the received updates and distributes the refined global model.
• Figure 3: Matching accuracy of strategy tuples under different LLMs.
• Figure 4: Federated evaluation number under different matching accuracies.
• Figure 5: Deployment time of different methods.