Evaluation of cluster K-Means as VM selection in dynamic VM consolidation

The paper focused on the evaluation of K-Means as VM Selection in Dynamic VM Consolidation. The research has compared the performance of several numbers of cluster K-Means in VM selection with another method such as Minimum Migration Time (MMT) and Random Choice (RC). Several attributes such as MIPS VM and RAM VM have been used for clustering the VMs using K-Means. Besides that, in the overload detection mechanism that works before VM selection, we used Median Absolute Deviation (MAD) during evaluation. Moreover, workload PlanetLab and CloudSim were used in the evaluation process. The parameter measurements in this research are Energy Consumption, SLAV, SLATAH, and PDM. The evaluation results have shown the different number of cluster in K-Means has influenced performance of VM selection in dynamic VM consolidation. K-Means also able to reduce energy consumption in cloud data center compared with other VM Selection methods.