Virtual Machine Placement Strategy Using Clustering and Genetic Algorithm for increasing cloud performance and power saving

Cloud computing is one of the most critical technologies of the century. However, it faces many challenges in both the adoption and operation phases. In the bigger picture, resolutions leading to increasing performance and decreasing costs are highly valued. Cloud computing has two types of expenses, Initial costs and routine expenditures such as human resources, maintenance fees, and electricity bills which form a significant part of the total costs of cloud infrastructure. Improving efficiency using optimal resource usage and decreasing energy waste would be possible through correct resource allocation. Intelligent algorithms can achieve maximum efficiency by placing virtual machines in the minimum number of physical servers and using the resources of these servers to the best. The main goal of this research is to reduce electricity consumption and the cooling required for under-utilized servers while preventing network congestions. The proposed algorithm of this research uses a multi-objective genetic algorithm and clusters the virtual machines to reduce the genetic algorithm execution time. The evaluation of the proposed algorithm implementation indicates that in this method convergence time is faster than the algorithms that lack clustering. As the secondary objective, the proposed algorithm distributes network traffic between physical servers to reduce network bottlenecks.