Research On Ddos Mitigation Methods In Edge Computing Based On Game Theory

Edge Computing is a new paradigm that enables relatively resource-poor mobile devices and Internet of Things(IoT)devices to enhance their capabilities and execute data/computation-intensive applications,while cooperating with resource-rich web servers to enable ubiquitous computing.As an extension of the cloud,it provides more resources for applications with low latency requirements.However,the new paradigm will also bring new security issues.The notorious Distributed Denial-of-Service(DDoS)attack,as a topic that has affected cloud computing for many years,has some new features in latency-sensitive edge computing environments.At the same time,edge servers with relatively scarce computing resources are more vulnerable under DDoS.This requires strengthening research on related security issues.Aiming at the characteristics of DDoS on the edge server,to the best of author's knowledge,this thesis first proposed Edge DDoS Mitigation(EDM).From the perspective of high-speed links between edge servers,it modeled the system cost of DDoS in edge computing,based on which,there were an optimal model and an integer programming model,and the centralized optimal solution is NP-hard.This thesis modeled EDM as an optimization problem and proved its NP-hardness.To implement EDM effectively,this thesis proposed a novel game theoretical approach named EDMGame for mitigating Edge DDoS attacks.EDMGame formulated EDM as a potential game with Nash equilibrium,and utilized a distributed algorithm to find out the Nash equilibrium as the implementation for EDM.Through theoretical analysis and experimental evaluation,the results illustrated that the proposed approaches in this thesis can efficiently and effectively mitigate Edge DDoS attacks.