Entropy-aware Energy-efficient Virtual Machine Placement in Cloud Environments Using Type Information
Description
Virtual machine placement (VMP) falls in the category of NP-hard knapsack problems. To overcome the time complexity of the problem, the use of heuristic and metaheuristic methods has attracted the attention of many researchers. In this paper, for the first time, we use an entropy-based method for VMP. The proposed method tries to place the VMs on physical machines by leveraging type information to minimize entropy. For this purpose, we use one of the most common entropy criteria called the Gini coefficient. We then solve the multi-objective problem with the non-dominant genetic sorting algorithm (NSGA-III). The simulation results on the CloudSim simulator along with statistical analysis show that the entropy-based method has a significant improvement over the state-of-the-art methods in terms of performance criteria such as utilization, resource wastage, and energy consumption.
Files
Steps to reproduce
Please read the README.pdf file.