DLopC: Data Locality Independency-Aware VM Clustering in Cloud Computing

Cloud Computing (CC) is one of the most popular technologies of the modern era, which provides seamless connectivity and services to the end users as per their demands. In CC, using virtualization, efficient resource utilization can be achieved which in turn increases the performance of any implemented solution in this environment. However, in order to provide the services to the end users, there is an exponential increase in the size and number of servers within the cloud data centers (DCs) which raises the need of efficient VM management in modern DCs. However, most of the clustering techniques reported in the literature are centroid based in which the number of VM clusters are predefined, which makes them highly dependent on the location of the data. Moreover, these existing proposals emphasize on arranging the data in a spherical structure with equal number of VM clusters. To mitigate all these issues, in this paper, we propose Tukey?s-HSD Based Clustering (TBC), which provides high accuracy for clustering a set of VMs. In the proposed scheme by using the two key parameters-CPU and RAM utilization, VMs are clustered. VMs having different resource utilization are compared using the Normal Curve and Analysis of Variance (ANOVA) test along with the TBC test for evaluating the membership of all VMs. With respect to these metrics, the performance of the proposed scheme is found to be superior in comparison to the other state-of-the-art competing schemes of its category.