Software-defined networking (SDN) is an emerging technology, which provides network architecture that decouples the control plane from the data plane. This main characteristic of SDN are bringing several of advantages. Due to the centralized control the network becomes more dynamic, and the network resources are managed in more efficient and cost-effective manner.
Another technology that focuses the attention of both the industry and the academy and has a huge potential to be wildly used all over the world is Vehicular Ad Hoc Networks (VANET). It is based on Mobile Ad Hoc Networks (MANET), in which the nodes are considered to be vehicle instead of mobile devices. VANETs are the key components of the intelligent transport systems (ITSs), whose major aim is to improve road safety and to provide different applications to the drivers and the passengers.
One of the main objectives of this thesis is to investigate how these two technologies can be implemented together, in order to achieve improved network performance. We claim that VANET networks can benefit from using SDN controller. Due to the separation between the control and data planes in VANET, network intelligence can be logically centralized and the underlying network infrastructure can be decoupled from the applications.
The centralized control of SDN brings an immense number of advantages, but it also can become a single point of failure of the network. The entire network could be compromised if the controller is under attack and therefore the network security in SDN-based VANETs is a major concern. In order to address some major security aspects of the VANET scenario, we estimate how Denial of Service Attack (DoS) and the Distributed Denial of Service Attack (DDoS) can influence the performance of SDN-based VANET network. The main purpose of this work is to detect DDoS attack of User Datagram Protocol (UDP) packets in order to meet the needs of real-time services, such as accident prevention, traffic jam warning, or communication.
This diploma thesis designs and tests a DDoS detection algorithm for SDN-based VANET networks. The test scenarios include launching normal and DDoS attack traffic with spoofed source IP addresses. Based on traffic features, entropy is used to measure the degree of randomness of occurrence of destination IP address of the packets. The algorithm is implemented as a software module on the SDN controller, by the means of two additional functions for detection of DDoS attacks. Entropy is calculated within predefined window size to measure uncertainty in the coming packets. After that the result is compared to a predefined threshold in order to classify the traffic as normal or attack traffic.