Hierarchical distributed edge data aggregation and reporting method based on cluster center selection

Abstract

With the surge of IoT devices, sensors, and smart terminals has led to distributed data sources and vast volumes of data. These challenges traditional centralized networks and cloud computing architectures, which struggle with bandwidth, latency, and storage limitations. Consequently, decentralized edge computing is crucial, enabling data processing and analysis at the network's edge to alleviate data return pressure and enhance system response speed and reliability. However, traditional centralized data aggregation methods become inefficient in the face of massive data and computing resources, resulting in long transmission times and low processing efficiency. To address these issues, this paper presents a hierarchical distributed edge data aggregation reporting method based on cluster center selection (HDAR-CCS). This method employs a staged approach to distributed data aggregation, utilizing parallel processing at each stage to efficiently handle data from multiple edge data centers. Additionally, an optimal cluster center selection algorithm is proposed, integrating the distances between cluster centers and available network resources. By establishing a selection criterion based on these distances, we design an effective scheme for choosing initial and subsequent cluster centers. Experimental results demonstrate that our approach outperforms existing algorithms, effectively meeting the low latency, high bandwidth, and efficient processing needs of intelligent applications.