GrC-VMM: An intelligent framework for virtual machine migration optimization using granular computing

Abstract

Virtual Machine Migration (VMM) is a critical component in cloud computing environments, enabling dynamic resource management and system optimization. However, existing approaches often face challenges such as increased downtime, excessive resource consumption, and complex decision-making processes in heterogeneous environments. This paper presents a novel framework based on Granular Computing (GrC) principles to address these challenges through systematic VM categorization and prioritization. The proposed framework employs a three-stage approach: (1) feature extraction and granule formation, converting VM attributes such as workload, downtime sensitivity, and resource utilization into meaningful information granules; (2) granule-based decision rule generation using formal GrC methodologies; and (3) priority-based classification using weighted membership functions. Experimental evaluations conducted using CloudSim 5.0 demonstrate the framework’s effectiveness across multiple performance dimensions. The results show 92. 1% classification accuracy, 83. 7% resource utilization and reduced migration downtime of 1.9 s. The framework exhibits linear computational complexity O(N), confirming its scalability for large-scale deployments. Additionally, performance analysis under various workload patterns (resource-intensive, service-oriented, and mixed) validates the framework’s robustness and adaptability. These results suggest that the proposed GrC-based approach offers a promising solution to optimize VM migration in cloud environments while maintaining operational efficiency and service quality.