Engineering Simulated Evolution for Virtual Machine Assignment Problem

Cloud computing is a rapidly growing services business in today's IT market. Its sharp growth is producing many challenges for cloud managers. One primary concern is to efficiently manage the cloud resources, i.e., to maximize utilization of hardware with minimum power consumption. Virtual Machine (VM) consolidation is a very helpful approach to achieve these goals. In this context, we investigate the VM assignment problem. We describe the engineering of a nondeterministic iterative heuristic known as Simulated Evolution (SimE) to solve the well-known NP-hard problem of assigning VMs to hardware hosts. A 'goodness' function which is related to the target objective of the problem is defined. It guides the moves and helps traverse the search space in an intelligent manner. In the process of evolution, VMs with high goodness value have a smaller probability of getting perturbed, while those with lower goodness value may be reallocated via a compound move. Results are compared with those published in previous studies, and it is found that the proposed approach is efficient both in terms of solution quality and computational time demand.