Online and Collaboratively Mitigating Multi-Vector DDoS Attacks for Cloud-Edge Computing

Edge computing is witnessing a convergence of cloud data centers and edge clouds, thereby the large thereby intensifying the vulnerability of cloud services from multi-vector DDoS attacks. However, existing DDoS filtering approaches, characterized by independent offline decisions made by clouds, exhibit shortcomings in efficacy and real-time performance. This paper proposed an online collaborative mitigation framework for multi-vector DDoS attacks, which formulates the mitigation challenge as an Online Multi-dimensional Multiple-Choice Knap-sack Problem (O-MdMCKP). Further, the framework generates candidate filtering policies for each incoming attack flow and designs a policy selection algorithm by employing online analysis based on reservation functions, ensuring prompt and efficient filtering. Experimental results show the proposed algorithm outperforms other online benchmark methods.