Article Preview
TopIntroduction
Cloud computing is emerging as a new paradigm of large-scale distributed computing (Armbrust et. al, 2010). It provides on-demand computing resources i.e., CPU, memory, network etc., as a service to the client through internet (Carretero and Garcia Blas, 2014; Slabeva and Wozniak, 2010; Wind et al., 2011). Virtualization is the technology that makes cloud computing possible (Haro et al., 2012; Younge et al., 2011). It allows sharing of the physical resources. It is implemented through the hypervisor also known as virtual machine monitor which reside between hardware and OS. When a user request for the resources, hypervisor engenders VM according to the user requirements, bind service to the VM and then assign to the user. Although virtualization increases the resource utilization, but it brings the problem of resource leakage and load balancing which needs to be tackled. Virtualization interfuses with the migration enable the cloud provider to balance the load on the server and consolidate their computing needs to a minimum number of servers. VM migration is a process which allows seamless movement of VM from one physical machine (PM) to another (Falco et al., 2015; Cerroni and Esposito, 2016; Wasim et al., 2015). Migration process consumes some resources for processing the data. Hence, successive migrations degrade the performance of the PM and must be evaded.
After reviewing the theory of cloud computing it is found that electricity bill highly affects the provider revenue which mainly depends on the number of active servers and CPU utilization (Beloglazov et al., 2012). A Recent study (Mishra and Sahoo, 2011; Xu et al., 2012), says that non-uniform resource utilization is one of the main causes for increasing the number of active servers, which will result in higher power consumption. Since cloud resources are multidimensional i.e., CPU, memory, network, bandwidth etc., so there is a situation where resource in one dimension is over utilized while it is underutilized in another dimension. This situation is known as resource leakage which is a phenomenon of resource wastage. Hence, resource leakage, number of migrations and energy consumption are the three critical issues in cloud and must be resolved for optimizing the cloud services.
Cloud is a business model where provider wants to generate higher revenue with minimum investment and high user gratification. Three metrics resource utilization, energy consumption and service level agreement (SLA) violation contributes major role in generating the high revenue to the service provider. These three metrics can be tackled effectively by utilizing the physical resources uniformly. So prime goal of the cloud provider is to reduce the resource leakage which results in optimizing the resource utilization. In order to achieve this goal an efficient load balancing approach is required that dynamically place the VM on the suitable host.