Power-Efficient Virtual Machine Placement in Cloud Datacenters using Heuristic Assisted Enhanced Discrete Particle Swarm Optimization

Abstract

    The increase in cloud computing services and the large-scale construction of data centers led to excessive power consumption. Datacenters contain a large number of servers where the major power consumption takes place. An efficient virtual machine placement algorithm is substantial to attain energy consumption minimization and improve resource utilization through reducing the number of operating servers. In this paper, an enhanced discrete particle swarm optimization (EDPSO) is proposed. The enhancement of the discrete PSO algorithm is achieved through modifying the velocity update equation to bound the resultant particles and ensuring feasibility. Furthermore, EDPSO is assisted by two heuristic algorithms random first fit (RFF) and random best fit (RBF) to produce hydride algorithms termed RFF-EDPSO and RBF-EDPSO. The proposed algorithms are evaluated and compared with recent algorithms to minimize power consumption. Simulation results showed the effective performance of RFF-EDPSO and RBF-EDPSO in minimizing the number of operating servers.