USER CLUSTERING MODEL OF A FRAMEWORK BUILT ON A MICROSERVICE SOFTWARE PLATFORM

Abstract

The paper presents the results of developing and evaluating a user clustering model designed for a framework implemented on a microservice-based software platform. The proposed approach focuses on efficient user grouping as a prerequisite for optimizing resource allocation and load distribution in distributed computing environments. The user clustering model is based on the k-means clustering methodology and enables the formation of stable user groups within predefined clustering boundaries under the condition of coincidence between the user clustering center and the corresponding cluster center.

The model operates by analyzing user distribution within a designated clustering zone and assigning users to clusters of predefined sizes according to the criterion of minimal distance between user distribution centers and cluster centroids. This approach ensures consistent clustering results and allows the system to maintain balanced clusters that reflect real user behavior patterns. As a result, the proposed model supports an accurate estimation of the computational load generated by each user group, which is essential for effective planning and management of computing resources in microservice-oriented systems.

The experimental results obtained during the applicability assessment demonstrate that the developed model is suitable for grouping framework users within the specified cluster boundaries and distribution constraints. The clustering outcomes confirm the model’s ability to adapt to varying user densities while preserving the required cluster sizes and structural stability. This, in turn, enables further evaluation of computing capacity demands associated with each cluster.

The presented user clustering model can serve as a foundation for advanced resource management mechanisms, including the identification of available Fog computing devices and the dynamic migration of microservices based on cluster characteristics. By mapping user clusters to appropriate Fog nodes, the proposed approach supports the implementation of distributed fog-based dynamic computing, improves system scalability, and enhances the overall efficiency and responsiveness of microservice platforms operating in heterogeneous and geographically distributed environments.