The wide application of cloud computing technology makes network security challenges increasingly complex, and traditional single-layer protection strategies are difficult to cope with diverse network attacks. This study proposes a network security emergency response mechanism and data protection strategy based on multilayer protection for the increasingly complex network security threats in cloud computing environment. Methodologically, a Partially Observable Markov Decision Process (POMDP) model is constructed, combined with an attack defense tree for security strategy decision-making, and the defense strategy benefit is quantified by fuzzy hierarchical analysis. The experiments are validated using real cloud platform data, and the results show that: in the analysis of the attack gain matrix, the maximum attack gain value under the high-risk state reaches 14.12; after the implementation of the optimal defense strategy, the defense gain matrix shows that the maximum defense gain can reach 54.8, which is significantly higher than the attack gain; and the experiment of the temporal strategy proves that, when the defense period (3DT) is smaller than the attack period (5AT), the percentage of infected nodes accounts for ratio is only up to 34.75%, and the network system quickly tends to the steady state at t=29s. The conclusion shows that the multilevel protection mechanism proposed in this study can effectively identify the optimal defense strategy and improve the network security level in cloud computing environment, which provides theoretical basis and technical support for practical application.
