Dynamic Virtual Machine Consolidation in a Cloud Data Center Using Modified Water Wave Optimization.

Cloud computing is the on-demand availability of computing resources, especially storage and computing power. The Cloud Data Center (CDC) consumes a huge amount of electrical energy due to inefficient resource utilization and scheduling algorithms. This leads to issues like high energy consumption along with high maintenance costs and carbon emissions. Virtual machine consolidation (VMC) is one of the energy-saving approaches used in data centers. The VMC problem in CDCs with energy constraints is a strict NP-hard problem. In this paper, a dynamic VMC algorithm is proposed to address the issues discussed above. The proposed algorithm categorizes the servers into three categories such as underloaded, overloaded, and normally loaded based on the CPU load. It migrates a few virtual machines (VMs) from overloaded machines to the normally loaded machine(s) for load balancing and migrates all the VMs of underloaded machines to normally loaded servers to switch-off idle servers. Further, modified water wave optimization (MWWO) approach is used to find a suitable migration plan which will reduce the load on overloaded servers and increase the overall resource utilization. It has been known and proven that an overloaded host consumes more energy over some time than the normally utilized host. The experimental evaluation is done on CloudSim Plus which is a cloud simulation tool. We have used two real-time workloads for experiments to reveal the efficiency of the proposed MWWO-VMC algorithm. The simulation experimental results are compared with various heuristic and metaheuristic dynamic VM consolidation algorithms. In comparison to previous similar works, the MWWO-VMC algorithm is successful in reducing energy consumption and VM migrations, increasing resource utilization and maximizing the dormant servers. [ABSTRACT FROM AUTHOR]