Article Preview
Top1. Introduction
Recently, cloud data centers have been paid considerable attention in industrial as well as academic communities. Multiple Virtual Machines (VMs) are allowed for re-locating on a single Physical Machine (PM) based on resource virtualization. Nowadays, the leading enterprises, such as Amazon and Microsoft, employed data centers for providing applications, like scientific computations, storage of large data, and hosting several web services. To fix the security vulnerability, few of the data centers undergo maintenance process, and therefore the continuous services are not provided to VMs at the particular time slots (Wang, et al., 2017; Karthikeyan, et al., 2018). With the rapid growth of cloud services in large and small scale industries, the data center sizes increasing continuously. Thus, the data centers need several cooling devices to keep the data center at a particular temperature, resulting in increased Carbon di oxide (CO2) emission and energy consumption (He, et al., 2019; Han, et al., 2019). The physical resources cost of the data center is 15%, and the cost of energy consumption is 45% (Narantuya, et al., 2018; Zakarya, 2018). Therefore, the factors mentioned above are very important for identifying energy-based resource allocation at the cloud data center to make a green data center (Kansal & Chana, 2016; Sharma & Reddy, 2016).
Cloud computing is implemented based on the distributed services, and it provides the virtualized resources like parallel semantic computing, distributed semantic models construction, semantic ambiguity, etc. The semantic web is employed to convert the World Wide Web into a structured intelligent web system and provide services to the users. The semantic web helps to express the domain's knowledge and organize the Metadata in cloud computing. Moreover, it helps solve the big data distribution in cloud computing and encourages data sharing (Meshram, et al., 2016). With the expeditious growth of the cloud for computation, storage and networking, live migration of VMs has given rise to the need for effectively managing data centers (Singh, & Gupta, 2016; Patel, et al., 2019). VM is widely utilized for data center management. In addition, virtualization-based solutions are adopted for generating VMs based on users' requests for computing resources, storage space, and the network bandwidth (Medina & García, 2014; Osanaiye, et al., 2017). When multiple VMs are available in a single server, the virtualization may enhance the underloaded server's utilization, leading to less power consumption at the fewer servers (Satpathy, et al., 2018). However, the energy-efficient-based resource management for the virtualization data centers becomes an attractive research area (Zhang, et al., 2018; Han, et al., 2019). The migration of VMs from one data center to another has made it possible to conveniently maintain data centers without affecting much of the performance of VMs. In live migration, data from a physical machine (PM) is copied to destination PM in another data center while the VM continuously runs on the former PM. Once the data is copied, the VM is continued on the new PM. It ensures negligible downtime and thus achieves high performance. During the migration of VMs, clock synchronization becomes paramount. The clocks must be synchronized and at higher precision (Zhang, et al., 2019 ; Patel, et al., 2019).