A Modified Grey Wolf Optimization for Energy Efficiency and Resource Wastage Balancing in Cloud Data-Centers

Abstract

Virtual Machine Placement (VMP) process is one of the significant issues in Cloud data-centers. This operation can manage VMs to be placed on the best available Physical Machine (PM) and it has a notable impact on the performance, resource utilization, and power consumption of the data-centers. Meta-heuristics swarm intelligence methods are a common approach of solving these types of optimization problems. Grey Wolf Optimization (GWO) is a meta-heuristic algorithm that has proved its advantages. In this paper, we propose a Modified Grey Wolf Optimizer (MGWO) to balance power consumption and memory resource wastage using virtual machine placement. The suggested algorithm compared with two existing greedy algorithms, First Fit (FF) and First Fit Decreasing (FFD). The simulation results revealed the effectiveness of the proposed MGWO to provide a tradeoff between energy efficiency and memory resource wastage. The average of resource wastage of the MGWO is less than 10 percent, considering the tradeoff between energy consumption and the memory wastage parameter. In addition, the convergence test of the MGWO has been compared with a Genetic Algorithm (GA). The proposed solution is useful for defining some constraints for controlling resource wastage, such as cores and memory wastage.