Networks of IoT devices are becoming increasingly important, but these networks are prone to cybersecurity issues. This work provides a novel approach for safer IoT networks: swarm-based IoT cybersecurity penetration testing by other IoT devices in the same network. To test this scenario, a simulation environment including three different penetration testing algorithms was developed. A linear penetration testing algorithm mimics human penetration testing activities and is used with a single agent and with multiple agents. A swarm-based algorithm utilizing queues adds communication between agents. The third algorithm is a swarm algorithm that uses Particle Swarm Optimization (PSO), thus adding a nature-based approach. All three algorithms are used to find vulnerabilities in simulated IoT networks of two different sizes. The networks are a smart home with 30 IoT devices and a smart building with 250 IoT devices. This study's results show the superiority of multi-agent approaches over linear, single-agent approaches to detecting unique vulnerabilities in a network. The swarm algorithms, which used communication between agents, outperformed the multi-agent approach with no communication. Additionally, the swarm algorithm utilizing queues demonstrated faster detection of vulnerabilities than the PSO algorithm. However, over time, the PSO outperformed the queue-based algorithm on the smart home scale. The smart building scale also provided faster detection for the queue-based algorithm than for the PSO. However, the PSO approach again provides better results over time and uses less computation time and memory resources.
| Identifer | oai:union.ndltd.org:ucf.edu/oai:stars.library.ucf.edu:etd2020-2762 |
| Date | 15 August 2023 |
| Creators | Schiller, Thomas |
| Publisher | STARS |
| Source Sets | University of Central Florida |
| Language | English |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Electronic Theses and Dissertations, 2020- |
Page generated in 0.0107 seconds