K-Means Clustering Based VM Placement Using MAD and IQR

Abstract

Cloud Data Centers (CDCs) have been developed into a virtual computing platform for businesses. Nevertheless, CDCs require significant power, which is essential for processor speed, particularly for Internet of Things (IoT) activities. Despite the existence of a significant amount of research in the green allocation of resource methodologies, it has been carried out to minimize the usage of the CDCs. Traditional systems mainly seek to minimize the number of physical machines (PMs) and rarely tackle the problems of overload and energy efficiency of the virtual machines (VMs) regulations concurrently. Furthermore, present systems cannot evaluate and redirect traffic from relevant sources to maximize the quality of service (QoS) supplied by CDCs. To improve energy saving, we attempt to enhance the adaptive four thresholds energy-aware framework for VM deployment energy efficiency (AFED-EF) scheme to improve energy savings. That is a unique adaptive energy-aware VM allocation and deployment technique for different applications to address these issues. We conducted a comprehensive exploratory program utilizing an authentic workload of over a million Planet Lab VMs. The research results demonstrate that our modified approach outperforms the AFED-EF and other existing traditional approaches, such as median absolute deviation (MAD), interquartile range (IQR), and overload detection using exponentially weighted moving average.