Explore Clustering Algorithms for Server Placement in Edge Computing

The surge in mobile device usage and the growing demand for rapid data processing have spurred the adoption of edge computing solutions. This paper introduces an innovative data clustering method aimed at optimizing the deployment of edge servers to improve processing efficiency, using Shanghai taxi booking data as a practical illustration. Our goal is to identify the best locations for edge servers, treating each taxi as an individual edge device. These servers will handle incoming requests and route them to the appropriate edge devices within their assigned groups. The primary objective is to minimize server count while ensuring the clustering algorithm's quality and accuracy. To achieve this, we explore various clustering algorithms, including Balanced Iterative Reducing and Clustering using Hierarchies, KMEANS, Ordering Points to Identify the Clustering Structure, and Density-Based Spatial Clustering of Applications. We utilize the Silhouette method to determine the optimal number of clusters. Our results reveal that the Balanced Iterative Reducing and Clustering using Hierarchies algorithm produces the most favorable outcomes with the optimal cluster count. This study presents a promising avenue for implementing an edge computing framework to enhance processing efficiency in urban environments.