Cloud Computing Based Packet Intrusion Detection for Hyper Fusion Network Storage

With hyper-fusion network storage becoming a rapidly growing field, it is attracting users from all walks of life, including businesses, government departments, and social networks. The implications of network intrusion on cloud computing are becoming increasingly dire. A design method based on Convolutional Neural Network (CNN) hypergraph feature reduction is proposed to improve the precision and calculation efficiency of network intrusion detection systems. To achieve data reduction, the optimal feature subset representation form can be optimized using the minimum distance measurement of the hypergraph structure and represented as a more significant form by the high order relation between the entities in the real world; secondly, the optimal feature subset is identified using the hypergraph helly feature recursion method, and intrusion detection is classified using the residual error-based CNN. Finally, experimental comparisons on the KDD CUP 1999 benchmark test set proved that the proposed method outperforms the industry standard according to detection precision, recall rate, and stability index.