DDoS online detection method in edge computing environment based on Attention-1D-CNN

The invention discloses a DDoS (Distributed Denial of Service) online detection method in an edge computing environment based on Attention-1D-CNN (Convolutional Neural Network), which comprises the following steps of: firstly, classifying traffic in a communication link according to different unloading tasks, so that the unloading safety of computing tasks in the whole link is not influenced when a single task is attacked; extracting attribute values of the traffic under the same task and carrying out normalization processing; secondly, the processed data are input into Attention-1D-CNN, channel Attention and space Attention are used for learning and detecting the contribution degree of data features to DDoS detection, redundant information lower than a feature threshold value is removed through a screening function, the complexity of the model learning process is reduced, and the model is rapidly converged; a simulation result shows that the detection accuracy of the OA1C model reaches up to 99.73% under the