Dynamic Federated Data Integration and Iterative Pipelines for Scalable E-Commerce Analytics Using Hybrid Cloud and Edge Computing

Abstract

Background information: IoT systems create a big stream of data with exponential growth. These make traditional centralized systems have difficulty processing their huge streams. Decentralized, fault tolerant, and scalable analytics in IoT could be supported using federated learning, resilient workflow orchestration, and hybrid task allocation. Interlacing them guarantees real time, secure computing, and makes optimum resource exploitation feasible across diverse IoT scenarios. Objectives: This paper assesses the feasibility of federated learning, resilient workflow orchestration, and hybrid task allocation for IoT analytics. The goals are to reduce latency, improve cost-effectiveness, enhance model accuracy, and optimize resource usage. The framework is designed to be scalable, reliable, and adaptable to various IoT applications. Methods: The methodology will integrate federated learning into privacy-preserving analytics, resilient workflows for fault-tolerant task execution, and hybrid task allocation to support optimized resource management. Performance metrics such as latency, cost efficiency, model accuracy, and resource utilization would then be assessed across individual and combined configurations to validate the effectiveness of the framework. Results: The results for the Full Model were outstanding at 18.4 ms latency, 92.5% cost efficiency, 94.8% model accuracy, and 95.1% resource utilization. The individual methods performed moderately better, and blended setups exhibit considerable synergies, establishing the soundness and scalability of the framework in real-time IoT analytics. Conclusion: The proposed methodology outperforms the traditional approaches in guaranteed safety, scalability, and efficiency of IoT analytics. Following enhancements will be focused on adaptive federated learning, advanced fault recovery, and support for emerging applications such as Internet of Things-based autonomous vehicles. Such enhancements will further strengthen this framework's applicability and scalability in dynamic environments.