A Networked System Dependability Validation Framework Using Physical and Virtual Nodes

Department

Program

Citation of Original Publication

Mehjabin, Suhee Sanjana, Ali Tekeoglu, Mohamed Younis, Mohammad Ebrahimabadi, Rahul Chandran, Tamim Sookoor, and Naghmeh Karimi. “A Networked System Dependability Validation Framework Using Physical and Virtual Nodes.” IEEE Access 11 (2023): 127242–54. https://doi.org/10.1109/ACCESS.2023.3330688.

Rights

This item is likely protected under Title 17 of the U.S. Copyright Law. Unless on a Creative Commons license, for uses protected by Copyright Law, contact the copyright holder or the author.
CC BY 4.0 DEED Attribution 4.0 International

Subjects

Abstract

Emerging applications in the context of smart cities pursue a decentralized design that often involves numerous networked components. To validate such a design, the scientific community has resorted to software based simulators to find a way around the complexity of building large scale physical network test-beds. Network Simulator-3 (ns-3) is one of the most popular platforms for this purpose where communication-related performance metrics, e.g., latency and throughput, can be evaluated. Yet, concerns exist about the viability of such a simulated approach when assessing dependability metrics, e.g., trust and resilience to cyberattacks, since the misbehavior is mainly defined by the evaluator. Incorporating physical nodes within the simulated network would be advantageous in that regard. Advances have been made to connect network simulators, e.g., ns-3, to virtual machines to emulate communication with real devices. However, all efforts in the literature so far have been limited to a single physical host. This paper presents a framework where many external physical devices can act as a part of the ns-3 simulator and interact seamlessly with the nodes within the simulated network via Docker containers. Hence, our framework enables scalable and cost effective experimentation to validate dependability metrics like fault-tolerance and attack resilience. We demonstrate the utility of the proposed framework in evaluating performance under a set of attack scenarios.