2018/12/10 by Akash Raj Narayanadoss, Narayanadoss, Akash Raj, Tram Truong-Huu +5
Computer Science · Engineering · #Network Security and Intrusion Detection #Internet Traffic Analysis and Secure E-voting #Vehicular Ad Hoc Networks (VANETs)
paper · pdf · doi:10.48550/arxiv.1812.03639
Recent developments in intelligent transport systems (ITS) based on smart\nmobility significantly improves safety and security over roads and highways.\nITS networks are comprised of the Internet-connected vehicles (mobile nodes),\nroadside units (RSU), cellular base stations and conventional core network\nrouters to create a complete data transmission platform that provides real-time\ntraffic information and enable prediction of future traffic conditions.\nHowever, the heterogeneity and complexity of the underlying ITS networks raise\nnew challenges in intrusion prevention of mobile network nodes and detection of\nsecurity attacks due to such highly vulnerable mobile nodes. In this paper, we\nconsider a new type of security attack referred to as crossfire attack, which\ninvolves a large number of compromised nodes that generate low-intensity\ntraffic in a temporally coordinated fashion such that target links or hosts\n(victims) are disconnected from the rest of the network. Detection of such\nattacks is challenging since the attacking traffic flows are indistinguishable\nfrom the legitimate flows. With the support of software-defined networking that\nenables dynamic network monitoring and traffic characteristic extraction, we\ndevelop a machine learning model that can learn the temporal correlation among\ntraffic flows traversing in the ITS network, thus differentiating legitimate\nflows from coordinated attacking flows. We use different deep learning\nalgorithms to train the model and study the performance using Mininet-WiFi\nemulation platform. The results show that our approach achieves a detection\naccuracy of at least 80%.\n