Attack Intent Recognition And Anti-attack Mechanism In Edge Computing Environment

Edge computing,as a strategy to alleviate resource congestion,has gradually become a new paradigm to meet the needs of the Internet of Things and local computing.Compared with cloud computing,edge computing migrates data computing or storage to the edge of the network,which can effectively reduce the transmission delay between edge or cloud servers and users,and avoid traffic peaks in the network.However,the security of terminal-edge-cloud devices themselves is still a problem that cannot be ignored.Terminal-edge devices are mostly resource-constrained devices,lacking protection measures like cloud computing devices,resulting in weaker defense performance of terminal-edge devices and a higher possibility of being attacked by malicious devices.Its distributed characteristics make it impossible to directly apply traditional centralized security mechanisms to the edge computing architecture.Therefore,this article focuses on devices security and users privacy issues,focusing on the identification of attack intentions and anti-attack methods in the edge computing environment to ensure the secure interaction between terminal-edge-cloud devices.First,this paper proposes an attack intention identification method based on partially observable Markov decision(POMDP),which comprehensively considers the connection between sequential decision and historical state,through the identification of device attack behavior and system state mining attack The real purpose behind the attacker lays the foundation for predicting follow-up attacks and making corresponding defensive measures.By introducing a dynamic decision-making network,the intention recognition environment is modeled as a limited and sequential decision-making process,and the POMDP mathematical framework is used to analyze the transition process between different states and predict the next attack behavior and intention.Then,combined with the state abstract technology,an attack intent recognition(DQAR)algorithm is designed based on deep Q learning,the accuracy,detection rate and false positive rate of different attack intent recognition algorithms were evaluated and compared through simulation experiments.The results show that the algorithm proposed in this paper has a good performance in the accuracy of intention recognition.Secondly,this paper designs an anti-attack method based on mean field game to minimize the damage of malicious attacks by seeking the optimal anti-attack strategy for large-scale terminal-edge-cloud devices.First,the offensive and defensive game process of terminal-edge-cloud devices is designed to seek Nash equilibrium strategy and the return and benefits of both offense and defense.Then combined with the mean field game to design the anti-attack model,the security defense problem of large-scale terminal-edge-cloud equipment is transformed into the mean field countermeasure problem,and the self-organizing neural network is used to approximate the mean field coupling equations.On this basis,a distributed artificial intelligence-driven security defense decision(AMSD)algorithm is designed to obtain the optimal solution for device security interaction,thereby improving the attack resistance of terminal-edge-cloud devices.Finally,the effectiveness of the self-organizing neural network is verified by numerical simulation,and the initial number of terminal-edge-cloud devices and the number of iterations of different security defense algorithms are evaluated.The results show that the AMSD algorithm is more suitable for edge computing architectures with large-scale terminal-edge-cloud devices.