Multi-resource balance Virtual Machine Consolidation Based on Extreme Learning Machine and Improved Grey Wolf Optimizer

In today’s rapidly evolving computing landscape, Cloud Data Centers (CDCs) play a crucial role in reshaping IT infrastructure deployment and management. Despite widespread adoption, many CDCs operate below optimal capacity, leading to higher power consumption and operating expenses. The ongoing challenge is to efficiently balance resource utilization, conserve energy, and ensure a high standard of Quality of Service (QoS). Current research often focuses on minimizing power usage through aggressive VM consolidation, which can result in SLA breaches and resource wastage due to unusable resource fragmentation. This study addresses the challenge by introducing the EML_IGWO algorithm, which combines the predictive capabilities of the Extreme Learning Machine (ELM) algorithm for forecasting host states, a multi-resource balance index for effective management of multidimensional resources, and the Improved Grey Wolf Optimizer (IGWO) for adeptly selecting the equilibrium solution from the Pareto frontier. The integrated approach aims to optimize both resource utilization and energy consumption, striking a suitable compromise between the conflicting objectives of SLA violation and power consumption. Experimental results, based on a real dataset in CloudSim and benchmark comparisons, highlight the algorithm’s outstanding performance across all metrics.