Grid systems and cloud servers are two distributed networks that deliver computing resources (e.g., file storages) to users' services via a large and often global network of computers. Virtualization technology can enhance the efficiency of these networks by dedicating the available resources to multiple execution environments. This chapter describes applications of virtualization technology in grid systems and cloud servers. It presents different aspects of virtualized networks in systematic and teaching issues. Virtual machine abstraction virtualizes high-performance computing environments to increase the service quality. Besides, grid virtualization engine and virtual clusters are used in grid systems to accomplish users' services in virtualized environments, efficiently. The chapter, also, explains various virtualization technologies in cloud severs. The evaluation results analyze performance rate of the high-performance computing and virtualized grid systems in terms of bandwidth, latency, number of nodes, and throughput.
Top1. Introduction
Virtualization technology is executed by a process unit (e.g., a single program and an operating system) inside a program environment, namely jail or sandbox, running in a physical machine, namely hosting machine. A powerful hosting machine can be used to provide a set of the virtual machines (VMs) interconnected by one or multiple virtual networks. A virtual network scenario emulates behaviors of the same scenario implemented with real computer systems. The main advantage of virtualization technology is that the main processes running in virtual machines behave, almost, truly as they are running on a real environment (Uhlig et al., 2005; Kim & Forsythe, 2010; Sahoo, Mohapatra, & Lath, 2010; Wang, Iyer, Dutta, Rouskas, & Baldine, 2013). This approach can be used in computer networks (e.g., grid systems and cloud servers) to reduce equipment and management costs compared to real scenarios. In this case, the hosting machine is used to implement the entire network to save financial costs of all the real equipments and infrastructure (e.g., wire and hubs) (Adabala et al., 2005; Di Costanzo, De Assuncao, & Buyya, 2009; Liang & Yu, 2015; Chen, Zhang, Hu, Taleb, & Sheng, 2015; Han, Gopalakrishnan, Ji, & Lee, 2015). Figure 1 illustrates the main elements of a virtual network laboratory: backbone, headquarters, regions, and sites. Backbone is a network to transport all the traffic among headquarters and regions. Headquarters are the central sites that involve main organization servers and applications. Regions contain one or more sites to manage the activity of desirable organization. Finally, sites indicate different offices of the organization and its end-users (Galán, Fernández, Ruiz, Walid, & de Miguel, 2004).
Figure 1. A schematic of virtual network laboratory
This chapter describes various applications of virtualization technology in grid systems and cloud servers. The chapter, initially, focuses attention on network virtualization, virtualized projects of computing systems, and virtualization technology in teaching. Then, it represents virtualization technologies for high-performance computing environments including virtual machine abstraction, feature comparison between virtualized environments, a framework for high-performance computing based on virtual technology, and performance evaluation of high-performance computing. Afterwards, grid architecture for liquid computing and services, grid virtualization engine, virtual clusters for grid computing, virtualization in In-VIGO, and evaluation results of virtualized grid systems are studied in Section 3. Section 4 discusses about virtualization technology in cloud servers that involves a sample architecture of cloud computing, types and roles of virtualization in cloud servers, from virtualization to private cloud services, and virtualization in high-performance IaaS cloud. Grid virtualization into cloud environment is discussed in Section 5. Finally, the chapter is concluded by Section 6.