AI-Based Security Parallel Path Selection Mechanism

Software-Defined Networks (SDN) and Network Function Virtualization (NFV) provide network functions virtually, resources, and controls in the fixth-generation (5G). However, SDN/NFV technology poses new security risks. Distributed Denial of Service (DDoS) attacks have emerged as one of the biggest security threats due to their simplicity, effectiveness, and low attack costs. Therefore, this article proposes a parallel path selection method to detect various types of DDoS attacks in SDN/NFV networks. Firstly, the AI models are virtualized as virtual network functions (VNF), which can provide security capabilities for the network. Different AI models are combined to form multiple sequential paths. Secondly, we design a heuristic algorithm based on sequential paths (HASP). This algorithm compares the VNFs' names and IDs of sequential paths, parallelizes the VNFs in the sequential paths, and constructs a set of parallel paths. Then, a Q-learning-based parallel path selection mechanism is designed (QPPS). This method can select the optimal parallel path with maximize path detection rate and minimize delay. Finally, the proposed QPPS method is validated in a prototype system. The experimental results demonstrate that the QPPS method provides optimal paths for different attack scenarios.