Network Virtualization and Emulation using Docker, OpenvSwitch and Mininet-based Link Emulation

• Main Author: Prabhu, Narendra
• Format: Others
• Published: ScholarWorks@UMass Amherst 2020
• Subjects: Network Virtualization; Docker; Containers; SDN; Emulator; Mobile area Network; Computer and Systems Architecture; Digital Communications and Networking
• Online Access: https://scholarworks.umass.edu/masters_theses_2/985; https://scholarworks.umass.edu/cgi/viewcontent.cgi?article=2016&context=masters_theses_2

With the advent of virtualization and artificial intelligence, research on networked systems has progressed substantially. As the technology progresses, we expect a boom in not only the systems research but also in the network of systems domain. It is paramount that we understand and develop methodologies to connect and communicate among the plethora of devices and systems that exist today. One such area is mobile ad-hoc and space communication, which further complicates the task of networking due to myriad of environmental and physical conditions. Developing and testing such systems is an important step considering the large investment required to build such gigantic communication arrangements. We address two important aspects of network emulation in this work. We propose a network emulation framework, which emulates the functioning of a hierarchical software defined network. One such use-case is described using a mobile ad-hoc network (MANET) topology within a single system by leveraging contemporary network virtualization technologies. We present various aspects of the network, such as the dynamic communication in the software domain and provide a novel approach to build upon existing emulation techniques. The second part of the thesis presents a dynamic network link emulator. This emulator enables suitable link property re-configurations such as bandwidth, delay and packet loss for networked systems using simulation software. We characterize the results of tests for the link emulation using a hardware and software testbed. Through this thesis, we aim to make a small yet crucial contribution to the niche area of software defined networks.