Paper in a peer-reviewed national conference
Denial-of-service test-bed for distributed location proof system
Security, Location Proof, Denial-of-Service, Distributed Denial-of-Service, Software-Defined Networking, Deep Learning
Pedro Teixeira, Samih Eisa, Miguel L. Pardal
INForum. Lisboa, Portugal. 2021
Location proof systems use many smart devices scattered across different geographic areas to provide witnessed proof of location to enable secure location-based services. However, Denial-of-Service (DoS) attacks can affect the system by slowing or shutting down the smart devices, making them inaccessible to its intended users. These attacks can happen even if the network devices are authenticated and are using encrypted communications. This is because DoS is a distinct class of attack that targets the availability of the system and requires different security solutions that are hard to deploy and test. In this work we provide a Mininet-based test-bed for an example Internet of Things system. The test-bed can emulate all the network nodes across a city, including sensors, servers and routers, and can generate both regular and DoS traffic. Using it, we can experiment with DoS detection and mitigation techniques before the application is deployed. We have run experiments with simple threshold analysis and also with sophisticated techniques based on Deep Learning. The end-goal is to deliver systems that are protected and stable even during DoS attacks.